{
    "exported_at": "2026-11-17 19:11:18",
    "count": 44,
    "records": [
        {
            "id": 25,
            "title": "Deep Learning-Based Breast Cancer Detection with IoMT",
            "title_ar": "",
            "title_en": "Deep Learning-Based Breast Cancer Detection with IoMT",
            "author": "Bo Li",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Auckland",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "New Zealand",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "transformers",
            "categories": "deep-learning; computer-vision; generative-ai; medical-ai; cnn; transformers; gan; medical-imaging",
            "keywords": "",
            "abstract": "Breast cancer remains a leading cause of cancer-related mortality among women worldwide, which demands innovations in early detection and diagnostic efficiency. Existing approaches often neglect peri-tumoral contextual features, rely on limited or unrealistic synthetic data, and impose latency bottlenecks in distributed healthcare settings. We propose a reserved-space strategy for Region of Interest (ROI) extraction that achieves 97.5% True Positive Rate (TPR) while preserving contextual features, thereby capturing diagnostically critical tumour margins that rigid bounding boxes overlook. To address data scarcity, we develop an enhanced Generative Adversarial Network (GAN) model incorporating Efficient Multi-scale Attention (EMA) modules and Hinge loss, which generates high-fidelity (15.2 Fréchet Inception Distance (FID)) ROI images that maintain anatomical plausibility in lesions and microcalcification diversity, resolving the over smoothing and pathological irrelevance of conventional GAN architectures. This GAN-augmented dataset improves classification AUC by 10% compared to traditional augmentation methods. Through systematic comparison, we demonstrate that the Swin Transformer outperforms the Convolutional Neural Network (CNN) and standard Vision Transformers (ViT) models in mammography classification with 86.8% Area Under the Curve (AUC). Finally, we design a distributed edge-fog-cloud Internet of Medical Things (IoMT) framework that implements our workflow to mitigate computational bottlenecks in clinical deployment.",
            "source_url": "https:\/\/openalex.org\/W7135345431",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "ResearchSpace (University of Auckland)",
            "source_record_id": "W7135345431",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:46",
            "updated_at": "2026-11-15 16:52:46"
        },
        {
            "id": 92,
            "title": "Explainable Artificial Intelligence in the Image Domain and its Applications to the Medical Field",
            "title_ar": "",
            "title_en": "Explainable Artificial Intelligence in the Image Domain and its Applications to the Medical Field",
            "author": "Mirtha Lucas",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "DePaul University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; responsible-ai; self-semi-supervised; predictive-analytics; cnn; gan; medical-imaging; xai; fairness",
            "keywords": "",
            "abstract": "This dissertation investigates the development of Explainable Artificial Intelligence (XAI) methods for deep learning models in the image domain, with a particular focus on medical imaging applications. Although neural networks achieve high predictive performance, their lack of interpretability limits their adoption in critical domains such as healthcare, where transparency and trust are essential. This work addresses this challenge by proposing novel approaches that improve the interpretability and reliability of model predictions. A primary contribution is the introduction of Riemann–Stieltjes Integrated Grad-CAM (RSI Grad-CAM), a gradient-based attribution method that generates more relevant and spatially localized saliency maps. The method is evaluated using quantitative metrics and statistical analysis, demonstrating improved performance over established techniques such as Grad-CAM and Integrated Grad-CAM. A secondary contribution is the analysis of the interpretability of intermediate layers in convolutional neural networks, demonstrating that clinically meaningful features, such as spiculation in lung nodules, can be identified within specific layers of the model. A third contribution is the development of a human-centered framework based on a Variational Autoencoder-Auxiliary Classifier Wasserstein Generative Adversarial Network (VAE-AC-WGAN), which generates contrastive explanations aligned with clinical reasoning. Experimental results demonstrate that the proposed method produces higher-quality reconstructions and more informative anomaly maps than baseline approaches (e.g., medXGAN), as reflected in improved structural similarity (SSIM) and classifier-based evaluation metrics (AUROC). Furthermore, human evaluation indicates that the generated explanations better capture clinically relevant features, such as spiculation, supporting their interpretability in medical imaging contexts. Overall, this work advances XAI by enhancing model transparency and supporting the trustworthy integration of AI systems into medical decision-making.",
            "source_url": "https:\/\/openalex.org\/W7172068913",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Digital Commons-DePaul (DePaul University)",
            "source_record_id": "W7172068913",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:46",
            "updated_at": "2026-11-15 16:52:46"
        },
        {
            "id": 161,
            "title": "Generative Quantum Machine Learning",
            "title_ar": "",
            "title_en": "Generative Quantum Machine Learning",
            "author": "Christa Zoufal",
            "supervisor": "",
            "degree_type": "phd",
            "degree_evidence": "نص السجل: «doctoral thesis»",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "",
            "governorate": "",
            "publication_year": 2021,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; generative-ai; quantum-ml; gan",
            "keywords": "",
            "abstract": "The goal of generative machine learning is to model the probability distribution underlying a given data set. This probability distribution helps to characterize the generation process of the data samples. While classical generative machine learning is solely based on classical resources, generative quantum machine learning can also employ quantum resources - such as parameterized quantum channels and quantum operators - to learn and sample from the probability model of interest. Applications of generative (quantum) models are multifaceted. The trained model can generate new samples that are compatible with the given data and extend the data set. Additionally, learning a model for the generation process of a data set may provide interesting information about the corresponding properties. With the help of quantum resources, the respective generative models have access to functions that are difficult to evaluate with a classical computer and may improve the performance or lead to new insights. Furthermore, generative quantum machine learning can be applied to efficient, approximate loading of classical data into a quantum state which may help to avoid - potentially exponentially - expensive, exact quantum data loading. The aim of this doctoral thesis is to develop new generative quantum machine learning algorithms, demonstrate their feasibility, and analyze their performance. Additionally, we outline their potential application to efficient, approximate quantum data loading. More specifically, we introduce a quantum generative adversarial network and a quantum Boltzmann machine implementation, both of which can be realized with parameterized quantum circuits. These algorithms are compatible with first-generation quantum hardware and, thus, enable us to study proof of concept implementations not only with numerical quantum simulations but also real quantum hardware available today.",
            "source_url": "https:\/\/openalex.org\/W3212842386",
            "pdf_url": "https:\/\/arxiv.org\/pdf\/2111.12738",
            "doi": "10.3929\/ethz-b-000514692",
            "handle": "",
            "repository_name": "arXiv (Cornell University)",
            "source_record_id": "W3212842386",
            "other_sources": "",
            "verification_status": "verified",
            "created_at": "2026-11-15 16:52:47",
            "updated_at": "2026-11-15 16:52:47"
        },
        {
            "id": 289,
            "title": "Natural Language Generation with Reinforcement Learning",
            "title_ar": "",
            "title_en": "Natural Language Generation with Reinforcement Learning",
            "author": "Thomas Scialom",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "France",
            "governorate": "",
            "publication_year": 2022,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "gan",
            "categories": "deep-learning; nlp; generative-ai; reinforcement-learning; responsible-ai; gan; fairness",
            "keywords": "",
            "abstract": "Natural Language Generation (NLG) is the subfield of Natural Language Processing, where the task is to produce natural language outputs. Despite the important progress fostered by the application of Deep Learning, generated texts are still inconsistent and contain factual inconsistencies. At the root cause, we argue in this thesis that deep learning models in NLG suffers from inherent flaws in algorithms, which limits their efficiency. At training time, the standard training strategy, Teacher Forcing, induces the so called exposure bias, a mismatch with inference time, where the errors accumulate. Moreover, NLG suffers from a second flaw: its the automatic evaluation does not reflect well human judgement. In this thesis, we explore how to improve both evaluation and training in NLG toward more reliable systems. In particular, we propose a Question Answering based metric. We show how this metric can be used as a reward in a Reinforcement Learning setup to improve NLG models. Toward this objective, we also explore learned rewards that are the discriminators, and introduce several new algorithms that benefit NLG during training and decoding times. In particular, we propose to combine Monte Carlo Tree Search with Generative Adversarial Networks, resulting in state-of-the-art models.",
            "source_url": "https:\/\/openalex.org\/W4392267362",
            "pdf_url": "https:\/\/hal.sorbonne-universite.fr\/tel-03951598v1\/document",
            "doi": "",
            "handle": "",
            "repository_name": "HAL",
            "source_record_id": "W4392267362",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:48",
            "updated_at": "2026-11-15 16:52:48"
        },
        {
            "id": 298,
            "title": "Large-Scale Traffic Flow Prediction Using Deep Learning in the Context of Smart Mobility",
            "title_ar": "",
            "title_en": "Large-Scale Traffic Flow Prediction Using Deep Learning in the Context of Smart Mobility",
            "author": "Arief Barkah Koesdwiady",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Waterloo",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Canada",
            "governorate": "",
            "publication_year": 2018,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; deep-learning; generative-ai; ai-applications; predictive-analytics; gan; ai-smart-cities",
            "keywords": "",
            "abstract": "Designing and developing a new generation of cities around the world (termed as smart cities) is fast becoming one of the ultimate solutions to overcome cities' problems such as population growth, pollution, energy crisis, and pressure demand on existing transportation infrastructure. One of the major aspects of a smart city is smart mobility. Smart mobility aims at improving transportation systems in several aspects: city logistics, info-mobility, and people-mobility. The emergence of the Internet of Car (IoC) phenomenon alongside with the development of Intelligent Transportation Systems (ITSs) opens some opportunities in improving the tra c management systems and assisting the travelers and authorities in their decision-making process. However, this has given rise to the generation of huge amount of data originated from human-device and device-device interaction. This is an opportunity and a challenge, and smart mobility will not meet its full potential unless valuable insights are extracted from these big data. Although the smart city environment and IoC allow for the generation and exchange of large amounts of data, there have not been yet well de ned and mature approaches for mining this wealth of information to bene t the drivers and traffic authorities. The main reason is most likely related to fundamental challenges in dealing with big data of various types and uncertain frequency coming from diverse sources. Mainly, the issues of types of data and uncertainty analysis in the predictions are indicated as the most challenging areas of study that have not been tackled yet. Important issues such as the nature of the data, i.e., stationary or non-stationary, and the prediction tasks, i.e., short-term or long-term, should also be taken into consideration. Based on this observation, a data-driven traffic flow prediction framework within the context of big data environment is proposed in this thesis. The main goal of this framework is to enhance the quality of traffic flow predictions, which can be used to assist travelers and traffic authorities in the decision-making process (whether for travel or management purposes). The proposed framework is focused around four main aspects that tackle major data-driven traffic flow prediction problems: the fusion of hard data for traffic flow prediction; the fusion of soft data for traffic flow prediction; prediction of non-stationary traffic flow; and prediction of multi-step traffic flow. All these aspects are investigated and formulated as computational based tools\/algorithms\/approaches adequately tailored to the nature of the data at hand. The first tool tackles the inherent big data problems and deals with the uncertainty in the prediction. It relies on the ability of deep learning approaches in handling huge amounts of data generated by a large-scale and complex transportation system with limited prior knowledge. Furthermore, motivated by the close correlation between road traffic and weather conditions, a novel deep-learning-based approach that predicts traffic flow by fusing the traffic history and weather data is proposed. The second tool fuses the streams of data (hard data) and event-based data (soft data) using Dempster Shafer Evidence Theory (DSET). One of the main features of the DSET is its ability to capture uncertainties in probabilities. Subsequently, an extension of DSET, namely Dempsters conditional rules for updating belief, is used to fuse traffic prediction beliefs coming from streams of data and event-based data sources. The third tool consists of a method to detect non-stationarities in the traffic flow and an algorithm to perform online adaptations of the tra c prediction model. The proposed detection approach is developed by monitoring the evolution of the spectral contents of the traffic flow. Furthermore, the approach is specfi cally developed to work in conjunction with state-of-the-art machine learning methods such as Deep Neural Network (DNN). By combining the power of frequency domain features and the known generalization capability and scalability of DNN in handling real-world data, it is expected that high prediction performances can be achieved. The last tool is developed to improve multi-step traffic flow prediction in the recursive and multi-output settings. In the recursive setting, an algorithm that augments the information about the current time-step is proposed. This algorithm is called Conditional Data as Demonstrator (C-DaD) and is an extension of an algorithm called Data as Demonstrator (DaD). Furthermore, in the multi-output setting, a novel approach of generating new history-future pairs of data that are aggregated with the original training data using Conditional Generative Adversarial Network (C-GAN) is developed. To demonstrate the capabilities of the proposed approaches, a series of experiments using arti cial and real-world data are conducted.",
            "source_url": "https:\/\/openalex.org\/W2802252725",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "UWSpace (University of Waterloo)",
            "source_record_id": "W2802252725",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:48",
            "updated_at": "2026-11-15 16:52:48"
        },
        {
            "id": 411,
            "title": "Deep Zero- and Few-shot Learning in Computer Vision",
            "title_ar": "",
            "title_en": "Deep Zero- and Few-shot Learning in Computer Vision",
            "author": "Hongguang Zhang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Australian National University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Australia",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "object-detection",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; transfer-learning; few-zero-shot; cnn; object-detection; video-analysis; gan",
            "keywords": "",
            "abstract": "Most CNN models rely on the large-scale annotated training data, and the performance turns to be low when the size of training data is limited, therefore numbers of large-scale datasets are proposed for vision tasks. Though training on large-scale dataset can significantly improve the performance, creating such datasets for novel scenarios is costly. Moreover, relying on the number of samples is not consistent with human's learning abilities, as humans have the ability to understand novel concepts from limited examples. Transfer learning is a machine learning topic that addresses the above limitations in current deep models, and it studies how to make machines exploit the acquired knowledge to solve new problems, thus learning by generalizing from existing to new concepts, a trait akin to humans. Empirically, transfer learning for computer vision problems includes the following subtopics: i) zero-shot learning, ii) few-shot learning, and iii) domain adaptation. In this thesis, we study zero-shot and few-shot learning to demonstrate how to learn better from limited training samples. Zero-shot learning requires learning a mapping that associates feature vectors extracted from images with semantic annotations that describe objects and\/or scenes of interest. Our first work addresses the weakness of linear mapping in previous zero-shot learning models by learning via non-linear kernelized projections between features and attribute spaces to promote accuracy. We propose an easy learning objective inspired by the Linear Discriminant Analysis method with an incoherence mechanism. The dominant challenge in the generalized zero-shot setting is the unbalanced data distribution during training and testing, which makes it difficult for the classifier to distinguish if a testing sample is from a seen or unseen class. Inspired by GAN-based models, we propose the model selection mechanism, in which we leverage two sources of datapoints to train the model selection classifier to recognize which test datapoints come from seen and which from unseen classes. This way, generalized zero-shot learning can be decoupled into two disjoint classification tasks. Few-shot learning is a variant of knowledge transfer which is formulated as a meta-learning process. Firstly we investigate so-called second-order pooling and Power Normalizations in fine-grained classification tasks to study a proper way of using co-occurrence statistics for such tasks. Following above experiments and analysis, we propose the novel second-order similarity network, which measures image relations via co-occurrence relation descriptors, and significantly improves the performance of few-shot learning. Hallucinating auxiliary samples for training is also a popular and effective way to improve the accuracy of few-shot learning. To this end, we propose a saliency-guided hallucination strategy, which employs a saliency detector to segment foregrounds and backgrounds from sampled images forming an episode, then we mix every foreground with all available backgrounds in convolutional feature space to synthesize descriptors representing objects in various scenes. Furthermore, we note that previous few-shot learning methods fail to address the scale- and location-mismatch between support-query objects, thus leading to the performance loss. Inspired by the spatial-pyramid matching, we propose a novel spatial- and scale-matching network to explicitly match the support-query pairs over different scales and locations. Last but not least, we consider a more challenging scenario, namely few-shot action recognition, to study knowledge transfer in the video domain. Apart from the spatial information, robust aggregation over the temporal mode is fundamental in this task. To address this issue, we propose a novel action relation network with permutation-invariant attentions and auxiliary self-supervision tasks.",
            "source_url": "https:\/\/openalex.org\/W3112219073",
            "pdf_url": "",
            "doi": "10.25911\/5f5a020a5c26f",
            "handle": "",
            "repository_name": "ANU Open Research (Australian National University)",
            "source_record_id": "W3112219073",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:49",
            "updated_at": "2026-11-15 16:52:49"
        },
        {
            "id": 486,
            "title": "Deep learning approaches for yield prediction and crop disease recognition",
            "title_ar": "",
            "title_en": "Deep learning approaches for yield prediction and crop disease recognition",
            "author": "Luning Bi",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Iowa State University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2022,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "transformers",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; ai-applications; predictive-analytics; optimization; cnn; rnn; transformers; gan; ai-agriculture",
            "keywords": "",
            "abstract": "The increase of the world population has brought significant challenges to the agriculture production system. Although mechanization has been realized in agriculture, many tasks (e.g., breeding, field inspection) are still labor-intensive and time-consuming. Therefore an automatic and intelligent solution is needed for the advancement of agricultural production. During this process, the biggest challenge is how to teach computers to understand the concepts in the real world. For example, an experienced expert can easily determine whether a plant is diseased or healthy. However, this may be challenging for the computer. Thus, the motivation of this dissertation study is to tackle these challenges in precision agriculture. This dissertation consists of four papers that propose different deep learning methods for the most challenging problems in agriculture. In the first paper, a genetic algorithm (GA)-assisted deep neural network was built for yield prediction using genetic information and environmental factors. In the global search phase, the GA was introduced to help determine the best initial weights of the neural network. In the local phase, random perturbation was used to avoid the local optimum. By using the proposed method, the root mean square error can be reduced by up to 10%. In the second paper, we proposed a generative adversarial network (GAN)-based approach to generate additional images for the classification of plant species and diseases using limited data. CNN was used as the basic network to classify species and diseases. GAN and label smoothing regularization (LSR) were combined to generate additional training images. Regular data augmentation techniques were also used to expand the dataset. The results showed that compared with using the real dataset only, the proposed method can improve the prediction accuracy by 6%. In the third paper, the potential of using satellite imagery for plant disease detection was explored. A gated recurrent units (GRU)-based model was presented for early detection of soybean sudden death syndrome (SDS) through time-series satellite imagery. The results showed that, compared to XGBoost and fully connected deep neural network (FCDNN), the GRU-based can improve the overall prediction accuracy by 7%. In addition, the proposed method can also be adapted to predict the future development of SDS. In the fourth paper, a transformer-based approach was proposed for soybean yield prediction using time-series camera images and seed treatments information. First, a vision transformer (ViT) base model was designed to extract features from the images. Then another transformer-based model was established to predict the yield using the time-series features. A case study was been conducted using a data set that was collected during the 2020 soybean-growing seasons in Canada. The experiment results show that compared to non-time series prediction and other baseline models, the proposed approach can reduce the mean squared error by 25%-40%. In conclusion, this dissertation aims to apply different state-of-art deep learning methods in agriculture. The study covers different topics, which range from yield prediction, species classification, to plant disease classification and prediction. At the model level, the application of linear models, tree-based methods, fully connected neural networks, convolutional neural networks, time-series models and transformers to different tasks have been investigated. In terms of the learning type, both unsupervised learning and supervised learning have been utilized. The experimental results have shown that appropriate deep learning methods can achieve better performance than traditional methods on specific tasks. Based on our work, more applications of deep learning techniques can be developed in the future.",
            "source_url": "https:\/\/openalex.org\/W7135318077",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Iowa State University Digital Repository (Iowa State University)",
            "source_record_id": "W7135318077",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:50",
            "updated_at": "2026-11-15 16:52:50"
        },
        {
            "id": 564,
            "title": "Self- and semi-supervised machine learning techniques for discovering novel ocular and neuroimaging biomarkers of Alzheimer's disease",
            "title_ar": "",
            "title_en": "Self- and semi-supervised machine learning techniques for discovering novel ocular and neuroimaging biomarkers of Alzheimer's disease",
            "author": "Charlene Z. L. Ong",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Nanyang Technological University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Singapore",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; responsible-ai; self-semi-supervised; predictive-analytics; cnn; gnn; multimodal; image-segmentation; gan; medical-imaging; xai",
            "keywords": "",
            "abstract": "There is increasing prevalence of Alzheimer’s disease worldwide. To manage the progression of Alzheimer’s disease, biomarkers can be utilized to identify patients with the disease or predict subsequent progression. While neuroimaging biomarkers have been long established to visualize structural changes associated with Alzheimer’s disease, there are other promising biomarkers such as retinal imaging biomarkers which offer the possibility of detecting early pathophysiological changes and are non-invasive and widely available. Nevertheless, the link between retinal imaging biomarkers to Alzheimer’s disease remains unclear and there is much difficulty in procuring datasets of patients for investigation, in particular longitudinal datasets. At the same time, advances in deep learning also offer much promise for furthering our understanding of the link between ophthalmic biomarkers and dementia. Self-supervised learning techniques offer potential for learning new biomarkers which might not have been defined previously. However, there is often a lack of sufficient ground truth annotations, hence hampering the application of deep learning models. In light of this, the work in this thesis aims to address these issues by exploring the use of self or semi-supervised learning approaches and leveraging on both neuroimaging and ophthalmic imaging. By utilizing unlabeled datasets, self or semi-supervised learning could potentially improve the model performance on a range of tasks, such as segmentation, disease classification, and prognosis prediction. More importantly, self-supervised learning techniques could uncover novel biomarkers in a data-driven manner. Some of these methods include generative adversarial networks, graph neural networks, and masked autoencoders, which are explored in this study for biomarker discovery. Firstly, the challenge of lack of sufficient ground truth annotation to develop models for retinal layer segmentation is investigated. Retinal layer segmentation is critical for quantifying the retinal layer thickness, which has been shown to be a useful diagnostic biomarker for age-related macular degeneration and a potential biomarker for Alzheimer’s disease. While there have been existing approaches developed which reduce annotation need, we propose to develop a deep learning approach using generative adversarial learning that utilizes a very limited number of labels for segmentation. We demonstrated that our approach showed comparable or superior performance to other state-of-the-art approaches, particularly with limited labels. In addition, we also evaluated our approach under various noisy conditions and showed it was robust in most noisy conditions. Secondly, current dementia diagnosis is based primarily on clinical assessment. Neuroimaging in the form of magnetic resonance imaging is used to provide additional granularity. It will hence be practical to explore neuroimaging biomarkers as a value add to clinical assessment, especially in a pre-dementia cohort. Our proposed approach uses a graph convolutional neural network to combine both clinical and non-clinical features such as neuroimaging for the classification of cognitive syndromes. It integrates a Cross Attention Module to highlight the salient features in the non-clinical features, attaining the highest accuracy and Matthew’s correlation coefficient compared to existing state-of-the-art methods. In addition, by investigating the feature interpretability via Integrated Gradients, several nonclinical features, such as cerebrospinal fluid volume, right hippocampal fissure volume, right frontal pole thickness, and left pallidum volume were identified as key features for the classification. Thirdly, we explore the utility of ophthalmic biomarkers as prognostic biomarkers for cognitive decline in Alzheimer’s disease patients. In parallel, we explore the possibility of combining both ophthalmic and neuroimaging features to form multimodal features. The application of these multimodal features is investigated in a clinical trial dataset consisting of participants with mild-to-moderate Alzheimer’s disease. To learn key feature representations associated with the ophthalmic images, we utilize self-supervised learning via masked autoencoders to extract feature representations. For neuroimaging, due to the availability of established neuroimaging software, we utilize standard preprocessing pipelines to obtain these neuroimaging features. We investigated the combination of these features using a graph convolutional network model for the prediction of cognitive decline in Alzheimer’s disease and showed initial promising results with the addition of optical coherence tomography data.",
            "source_url": "https:\/\/openalex.org\/W7127051058",
            "pdf_url": "",
            "doi": "10.32657\/10356\/205534",
            "handle": "",
            "repository_name": "DR-NTU (Nanyang Technological University)",
            "source_record_id": "W7127051058",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:50",
            "updated_at": "2026-11-15 16:52:50"
        },
        {
            "id": 568,
            "title": "Improving Human Behavior Understanding with Semi-supervised Learning",
            "title_ar": "",
            "title_en": "Improving Human Behavior Understanding with Semi-supervised Learning",
            "author": "Bruno Manuel Degardin",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Portugal",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "transformers",
            "categories": "machine-learning; deep-learning; computer-vision; nlp; generative-ai; responsible-ai; self-semi-supervised; cnn; transformers; multimodal; video-analysis; gan; diffusion; fairness",
            "keywords": "",
            "abstract": "Recent advancements in Computer Vision (CV) and Machine Learning (ML) have fundamentally revolutionized the field of human behaviour understanding, enabling intelligent systems to move beyond simple motion detection to the interpretation of complex, semantic actions from vast streams of visual data. This paradigm shift has been driven by the exponential growth of Deep Learning algorithms, particularly the evolution from 3D Convolutional Neural Networks (3D CNNs) to modern Vision Transformers (ViTs) , which have achieved remarkable success in modeling long-range spatiotemporal dependencies in RGB videos and multimodal signals. However, despite their impressive performance on controlled benchmarks, the deployment of these pixel-based models in real-world scenarios remains heavily constrained. Their reliance on high-dimensional visual data often incurs prohibitive computational costs that hinder real-time application on edge devices. Furthermore, these methods exhibit significant sensitivity to background clutter and viewpoint variations , and the processing of raw video raises critical privacy concerns by capturing unnecessary appearance details such as identity and clothing. Maximizing the potential of this modality presents a significant set of challenges that hinder widespread deployment. The heavy reliance on massive, high-quality annotated datasets creates a prohibitive bottleneck for supervised learning, while existing generative models used for data augmentation often lack physical plausibility or controllability. Furthermore, self-supervised methods borrowed from video domains frequently fail to capture the fine-grained dynamics of skeletal motion, and current evaluation metrics for synthetic data struggle to distinguish between realistic kinematics and artifacts. Addressing these limitations requires a fundamental shift toward more label-efficient, physically aware, and generative approaches to behaviour analysis. In this thesis, we proposed a comprehensive suite of novel solutions to mitigate these problems by exploiting semi-supervised and self-supervised learning paradigms, as well as advanced generative modeling. First, we addressed the limitations of temporal reasoning in Graph Convolutional Networks (GCNs) by proposing REGINA, a reasoning- augmented architecture. This method injects explicit temporal self-similarity cues into the network through a learnable temporal adjacency matrix, allowing the model to adaptively optimize its receptive fields to capture long-range dependencies that standard convolutions often miss. Addressing data scarcity, we proposed Kinetic-GAN, a generative adversarial network that scales action conditioning by an order of magnitude over prior works. By generating actions directly from a disentangled latent space using a mapping network and stochastic noise injection, it synthesizes diverse, high-fidelity global body movements across more than 100 action classes. We further advanced representation learning without labels through ATOM, a self-supervised strategy that leverages atomic motion representation learning. By decomposing actions into short, fine-grained primitives and training the backbone to recover their temporal order, ATOM learns robust features from the intrinsic structure of motion, transferring effectively to supervised tasks without the computational burden of heavy reconstruction heads. To bridge the gap between generation and recognition, we developed an intrinsic, kinematics-aware quality assessment framework. Recognizing that conventional distribution metrics fail to capture physical validity, this framework utilizes per-sample kinematic indicators within a weaklysupervised discriminator to iteratively filter synthetic samples, ensuring only physically plausible data is used to boost downstream performance. Finally, pushing the boundaries of controllable synthesis, we proposed GLDM, a Spatiotemporal Graph Latent Diffusion Model. GLDM projects skeleton sequences into a compact, graph-aligned latent space and performs diffusion with graph-aware updates. Crucially, by integrating cross-attention mechanisms, GLDM enables precise multi-modal conditioning, allowing generation to be guided not just by class labels but also by natural language prompts, achieving generalization beyond training data, unconstrained temporal flexibility and solving the mode collapse issues inherent to adversarial approaches. Results obtained by the proposed methods show that it is possible to improve HBU systems regarding label efficiency, generative controllability, and physical plausibility, while still obtaining state-of-the-art results. All proposed methods were evaluated on several large-scale benchmarks, demonstrating that robust behaviour understanding can be achieved with significantly reduced reliance on manual annotation.",
            "source_url": "https:\/\/openalex.org\/W7225277943",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)",
            "source_record_id": "W7225277943",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:50",
            "updated_at": "2026-11-15 16:52:50"
        },
        {
            "id": 590,
            "title": "Discriminator-driven diffusion generative adversarial network for synthetic medical image generation",
            "title_ar": "",
            "title_en": "Discriminator-driven diffusion generative adversarial network for synthetic medical image generation",
            "author": "Dennis Chia Yin Lim",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Australia",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "deep-learning; computer-vision; generative-ai; medical-ai; gan; medical-imaging",
            "keywords": "",
            "abstract": "This research project addresses the significant challenge of class imbalance in medical deep learning classification models, particularly when the abnormal class is underrepresented. Class imbalance can severely affect the classification performance of deep learning models, potentially leading to patient misdiagnosis and impeding clinician trust in AI adoption. In healthcare, class imbalance arises due to various factors, including the rarity of diseases or medical scenarios and data insufficiency for specific classes within large datasets. Existing solutions, such as traditional data augmentation and Generative Adversarial Networks (GANs) like DCGAN, as well as Conditional GANs like Pix2Pix and Stable Diffusion, have attempted to mitigate this issue, but with limitations.",
            "source_url": "https:\/\/openalex.org\/W7131907591",
            "pdf_url": "",
            "doi": "10.25916\/sut.31368169",
            "handle": "",
            "repository_name": "Swinburne figshare (Swinburne University of Technology)",
            "source_record_id": "W7131907591",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 591,
            "title": "Generative modelling under epistemic uncertainty",
            "title_ar": "",
            "title_en": "Generative modelling under epistemic uncertainty",
            "author": "Muhammad Mubashar",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Oxford Brookes University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United Kingdom",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "llm",
            "categories": "machine-learning; deep-learning; computer-vision; nlp; generative-ai; robotics; predictive-analytics; llm; gan; diffusion",
            "keywords": "",
            "abstract": "The deployment of Deep Learning in safety-critical domains is hindered by pathological overconfidence and an inability to distinguish between aleatoric uncertainty (data ambiguity) and epistemic uncertainty (lack of knowledge). This thesis addresses these limitations by establishing a rigorous framework for Random-Set Deep Learning, shifting from point-estimate probabilities to belief functions over the power set of outcomes. Enabled by a novel Budgeting strategy that ensures scalability, this framework allows models to explicitly represent ignorance. While we validate these principles through Random-Set Neural Networks (RS-NN) for classification and a Unified Evaluation Framework, the primary contribution of this work lies in re-imagining Generative AI under epistemic uncertainty. We introduce Random-Set Large Language Models (RS-LLMs), which predict belief functions over token sets to quantify second-order uncertainty, thereby providing a robust mechanism for hallucination detection. Furthermore, we propose Epistemic Generative Adversarial Networks (GANs) and Epistemic Diffusion Models. By modeling the uncertainty of the generation process itself via second-order distributions, these architectures significantly mitigate mode collapse and enhance sample diversity compared to standard baselines. Finally, we demonstrate the practical utility of these generative capabilities in Autonomous Driving. We introduce the ROAD-INTENT dataset and utilize Random-Set Vision Language Models (RS-VLMs) to automate the annotation of actor intent with calibrated uncertainty scores, paving the way for safer, self-aware AI systems.",
            "source_url": "https:\/\/openalex.org\/W7135169854",
            "pdf_url": "",
            "doi": "10.24384\/j47e-p788",
            "handle": "",
            "repository_name": "Radar (Oxford Brookes University)",
            "source_record_id": "W7135169854",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 592,
            "title": "Towards data efficient generative models",
            "title_ar": "",
            "title_en": "Towards data efficient generative models",
            "author": "Zhaoyu Zhang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Queen's University Belfast",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United Kingdom",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "gan",
            "categories": "machine-learning; computer-vision; generative-ai; gan; diffusion",
            "keywords": "",
            "abstract": "The past decade has seen significant advancements in generative models. As one of the most fundamental machine learning and computer vision tasks, generative models aim to produce realistic-looking synthetic samples by learning from real samples. However, training generative models from scratch commonly requires large amounts of data, which can be costly, time-consuming, or even impractical to collect and clean for specific domains. To address this, data-efficient generative models, which are designed to train with limited data from scratch, have gained significant attention. In this thesis, we focus on the theory and algorithms for training data-efficient generative models from scratch using two widely used generative models, i.e., generative adversarial networks (GANs) and diffusion models (DMs). For GANs, we address the challenges in data-efficient training from three different theoretical perspectives: the unfair min-max game, the overfitting of discriminator ($D$) and the leaking of augmentations problem. For DMs, we address the data efficiency challenge by better minimizing the total denoising score matching error during training. Firstly, we apply information game theory to analyze the GANs' min-max game and unveil a novel perspective on GAN training: the min-max game in existing GANs training is unfair. Based on this, we present a novel GAN called Information Gap GAN (IGGAN) to enhance the min-max game in GANs training. Specifically, IGGAN comprises a generator ($G$) and two discriminators ($D_1$ and $D_2$), each applying distinct data augmentation (DA) techniques. The information gap caused by various DA methods between $D_1$ and $D_2$ can alter the information received by each player in the min-max game, leading to incomplete information for all three players—$G$, $D_1$, and $D_2$—in IGGAN. As a result, IGGAN outperforms other existing methods, achieving better Inception Scores and lower Fréchet Inception Distances. Secondly we introduce a straightforward yet effective approach named Dual Adaptive Noise Injection (DANI) to mitigate the overfitting of $D$ in DE-GANs training. Specifically, DANI incorporates two adaptive strategies: adaptive injection probability and adaptive noise strength. By dynamically adjusting these factors, DANI effectively alleviates the overfitting of $D$. The strong performance and generalization ability of DANI demonstrate its effectiveness in improving GAN training under limited data conditions. Thirdly, we first illustrate that applying Data augmentation (DA) in DE-GANs can introduce out-of-distribution samples, leading to the undesirable leaking of augmentation issue in DE-GANs training. To address this, we present a straightforward yet powerful technique named Adaptive Negative Data Augmentation (ANDA) for DE-GANs. Specifically, ANDA enhances the augmented distribution of generated data by incorporating the augmented distribution of negative real data, which is generated by applying Negative Data Augmentation (NDA) to real data. This approach allows potential leaking samples to be presented as ``fake” instances to the discriminator in an adaptive manner, preventing the generator ($G$) from learning these samples and improving overall performance. We demonstrate that ANDA is effective in mitigating the leaking of augmentation issue. Finally, we focus on training diffusion models with limited data. We present a novel theoretical insight for diffusion models that two factors, i.e., the denoiser function hypothesis space and the number of training samples, can affect the denoising score matching error of all training samples. Based on this theoretical insight, it is clearly evident that the total denoising score matching error is hard to be minimized within the denoiser function hypothesis space of existing methods when training diffusion models with limited data. To address this, we design a new diffusion model with limited data called Limited Data Diffusion (LD-Diffusion), which consists of two main components: a compressing model and a novel mixed augmentation with fixed probability (MAFP) strategy. Comprehensive experiments show that LD-Diffusion greatly enhances the training of diffusion models with limited data, achieving state-of-the-art performance compared to other diffusion models.",
            "source_url": "https:\/\/openalex.org\/W7215036991",
            "pdf_url": "",
            "doi": "10.17034\/32826059",
            "handle": "",
            "repository_name": "Research Portal (Queen's University Belfast)",
            "source_record_id": "W7215036991",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 593,
            "title": "Generative Models from the Inside: a journey through properties and applications",
            "title_ar": "",
            "title_en": "Generative Models from the Inside: a journey through properties and applications",
            "author": "Barbara Toniella Corradini",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Florence",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Italy",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "llm",
            "categories": "deep-learning; computer-vision; nlp; generative-ai; transformers; image-segmentation; llm; gan; diffusion",
            "keywords": "",
            "abstract": "Generative Artificial Intelligence (Gen-AI) is a rapidly evolving field focused on developing models that can generate realistic and novel content, from text to images. Recent advancements such as ChatGPT and DALL-E have revolutionized domains like Natural Language Processing and Computer Vision by creating models capable of producing highly realistic and coherent outputs. This increase in generative capabilities has been driven by foundational models like Transformers, Generative Adversarial Networks (GANs), and Diffusion Models, which have expanded the boundaries of what machines can create. This thesis addresses open challenges in the field, including optimizing computational efficiency, analyzing the latent spaces of generative models, and extracting valuable insights from pre-trained models. The contributions of this work include the development of novel recursive architectures, the application of topological data analysis to better understand GAN training dynamics, and the use of diffusion models for training-free segmentation, highlighting the increasing versatility of generative models across diverse applications. The future of Gen-AI is promising, with numerous opportunities yet to be explored. While progress has been significant, much remains to be learned, and further exploration will be essential to fully realize the potential of these models. We hope that this thesis will provide a meaningful contribution to inspire a deeper understanding and development of generative modelling.",
            "source_url": "https:\/\/openalex.org\/W7213486992",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Florence Research (University of Florence)",
            "source_record_id": "W7213486992",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 595,
            "title": "Synthesizing High B-Value Diffusion-Weighted Imaging of the Prostate using Diffusion Models",
            "title_ar": "",
            "title_en": "Synthesizing High B-Value Diffusion-Weighted Imaging of the Prostate using Diffusion Models",
            "author": "Pedro Rodrigues Parracho",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Portugal",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "deep-learning; computer-vision; generative-ai; medical-ai; image-segmentation; gan; diffusion; medical-imaging",
            "keywords": "",
            "abstract": "Prostate cancer (PCa) remains among the most prevalent and deadly cancers among men worldwide, with rising incidence rates placing increasing pressure on healthcare systems. Magnetic Resonance Imaging (MRI), and more specifically, Diffusion-Weighted Imaging (DWI), has proven to be an ideal diagnostic tool in the early detection of PCa. However, acquisition of high b-value DWI, which is critical in detecting diffusion-restricted lesions, is challenging due to the prolonged scan times and susceptibility to noise and motion artifacts. This dissertation explores the clinical practicability of generating high b-value prostate DWI from low b-value prostate DWI inputs using generative artificial intelligence (AI) models with the aim of decreasing acquisition times and artifact susceptibility. Three generative models were tested: Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Denoising Diffusion Probabilistic Models (DDPMs). Each model was conditioned to take any low b-value DWI and generate a resulting DWI at any requested high b-value. Quantitative evaluation was conducted with image similarity metrics, comparing the generated DWI with the real acquired DWI. Diagnostic performance was also quantified with a deep learning lesion segmentation model in an attempt to determine diagnostic preservation in the generated DWI. The DDPM architecture was the most effective generative architecture, with the best performance in quantitative scores, in addition to high visual realism in maintaining diffusion effects. The generated images also preserved the majority of the diagnostic accuracy of the original high b-value scans, with preservation of lesion detection rates and localisation of lesions. The results verify that diffusion models are the most adequate architecture for low-to-high b-value DWI generation and can retain diagnostic performance. This study proved that generative AI can significantly enhance MRI workflows, reducing scan times, and facilitating scalable PCa screening programs, especially in resource-constrained environments, paving the way for more accessible diagnostic pathways.",
            "source_url": "https:\/\/openalex.org\/W7224044743",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT)",
            "source_record_id": "W7224044743",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 596,
            "title": "Non-Markovian Modelling of Limit Order Books with Generative Adversarial Networks",
            "title_ar": "",
            "title_en": "Non-Markovian Modelling of Limit Order Books with Generative Adversarial Networks",
            "author": "Navindu Subuddhi Nawurunna Palliyaguruge",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Tampere University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Finland",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; generative-ai; predictive-analytics; gan",
            "keywords": "",
            "abstract": "This thesis studies the use of Wasserstein GAN with Gradient Penalty and Temporal Attention augmented Bilinear Layers (TABL) to generate synthetic limit order book (LOB) data. The primary goal is to assess whether a non-Markovian conditioning of future snapshots on LOB history improves realism and mid-price movement prediction relative to a Markovian baseline. The proposed W-GAN models learn price dynamics implicitly from historical order flow, without relying on explicit mid-price movement labels. Results demonstrate that non-Markovian models consistently generate more realistic LOB snapshots and achieve higher F1 scores in mid-price movement classification than the baseline model. Among them, a history length n =30 performed the best in matching stylised facts and mid-price movements. They also incorporate LOB snapshots with empty queues in training, and learns to reproduce empty queues, but less frequently than in real data. Computational constraints required simplifying the model outputs, resulting in minor compromises to the quality of the generated samples. These findings highlight the benefits of non-Markovian conditioning in LOB generation and suggest promising directions for future research on model stability, scalability, and cross-asset generalisation.",
            "source_url": "https:\/\/openalex.org\/W7134871345",
            "pdf_url": "https:\/\/trepo.tuni.fi\/bitstream\/10024\/233187\/2\/NawurunnaPalliyagurugeNavindu.pdf",
            "doi": "",
            "handle": "",
            "repository_name": "Tampere University Institutional Repository (Tampere University)",
            "source_record_id": "W7134871345",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 597,
            "title": "Image restoration and representation with deep generative models",
            "title_ar": "",
            "title_en": "Image restoration and representation with deep generative models",
            "author": "Yufei Wang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Nanyang Technological University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Singapore",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "image-segmentation",
            "categories": "computer-vision; generative-ai; responsible-ai; ai-applications; image-classification; image-segmentation; gan; diffusion; fairness; ai-industry",
            "keywords": "",
            "abstract": "A fundamental challenge in the realm of computer vision lies in accurately modeling\/characterizing image distributions. For example, in high-level vision tasks, better representations of images in the latent space can significantly enhance downstream processes such as image classification and segmentation. Similarly, in image restoration, a more accurate model for the distributions of clean images conditioned on degraded images can yield results with better perceptual quality. However, modeling these distributions is highly challenging due to the high dimensionality of both images and their latent codes. In a relatively parallel research avenue, deep generative models have made remarkable strides, aiming to implicitly or explicitly model image distributions. These models offer potent tools for handling intricate distributions. However, directly applying these models is unfeasible and leads to suboptimal performances due to different domain priors and task formulations. The challenge of seamlessly and effectively integrating generative models into image restoration and representation tasks, along with incorporating task-specific priors, remains an open area for further exploration. This thesis focuses on exploring the potential of deep generative models in image restoration and representation tasks, spanning from high-level vision tasks, i.e., image classification, to low-level vision tasks, i.e., image compression and restoration. Firstly, for the image classification task, we introduce a novel variational inference framework to obtain a latent representation with enhanced generalization ability. We implicitly model the posterior distributions of images given their latent codes using a generative adversarial network to disentangle domain-invariant features from the provided training data. Subsequently, instead of implicitly modeling the posterior distributions, we explore the possibility of explicitly modeling them. Utilizing the power of Normalizing Flow, a generative model capable of obtaining the exact likelihood of a given sample, we apply it to the task of low-light image enhancement (LLIE), considering its inductive bias. While the flow-based model achieves promising results, the performance is constrained by the intrinsic limitations in the design of normalizing flow. Therefore, we delve into seamlessly integrating the prior knowledge of the LLIE task in raw image space into a diffusion framework to overcome the limitations in model designs. Although our diffusion-based solution has promising results, it still has two main limitations: first, even with the unique advantages of enhancement in raw space, the significant storage overhead of raw images hampers its application. Consequently, in the subsequent two works, we explore joint compression by accurately modeling the latent distributions of images and leveraging the capabilities of autoregressive models to further enhance coding efficiency. Second, while the proposed work shortens the inference path from tens or hundreds to only three steps compared with commonly used diffusion models, it still requires iterative evaluations and causes inference overhead. In the last work, we propose an acceleration strategy incorporating a distillation strategy with a novel self-consistency loss. We achieve SOTA performance in only one inference step in the super-resolution task, where the conditional distribution is relatively more complex. In summary, the thesis makes three primary contributions. First, we showcase the effectiveness and unique advantages of generative-based image restoration and representation methods. Second, we propose diverse approaches to seamlessly integrate the capabilities of deep generative models with the domain knowledge specific to image restoration and representation tasks. To validate the effectiveness of these proposed methods, extensive experiments are conducted across multiple datasets. The experimental results unequivocally demonstrate that our methods outperform prior state-of-the-art models. The efforts and achievements presented in this thesis not only underscore the practical capabilities of image restoration and representation techniques but also provide fundamental support for future research and applications within the industry.",
            "source_url": "https:\/\/openalex.org\/W4408105674",
            "pdf_url": "",
            "doi": "10.32657\/10356\/182543",
            "handle": "",
            "repository_name": "DR-NTU (Nanyang Technological University)",
            "source_record_id": "W4408105674",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 598,
            "title": "On large deviations in probabilistic deep learning and generative modeling",
            "title_ar": "",
            "title_en": "On large deviations in probabilistic deep learning and generative modeling",
            "author": "Nilsson, Viktor",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "KTH Royal Institute of Technology",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Sweden",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "llm",
            "categories": "machine-learning; deep-learning; computer-vision; nlp; generative-ai; medical-ai; responsible-ai; predictive-analytics; cnn; image-segmentation; llm; gan; diffusion; xai",
            "keywords": "",
            "abstract": "The topic of this thesis is the use of probabilistic methods in machine learning. These play a foundational role in motivating and defining machine learning algorithms, as well as in explaining why, and how well, the algorithms work. During the years in which this work was conducted (2020-2025), machine learning has gone from delivering a handful of impressive demonstrable results to becoming a staple of modern developed society, with products such as ChatGPT, etc., having capacity and applicability far beyond what anyone in the field expected ten years ago. Explaining why the modern methods work so well, despite their conceptual simplicity and elegance, requires both empirical and theoretical studies. This thesis has both parts, while the emphasis is on theory. The first part of the thesis, Papers A-C, concerns the implementation and analysis of novel methodologies in deep learning, whereas Papers D-F concern purely theoretical large deviations results for machine learning adjacent models. The main thread is the application of mathematical tools from probability theory and statistics, such as the theory of large deviations and empirical process theory, to the understanding and improvement of machine learning methodology. In Paper A, which presents the most applied direction pursued in the thesis, a deep probabilistic network model is applied to a task from the field of clinical radiation therapy, namely dose prediction, where a value of radiation dose target shall be assigned to each pixel\/voxel of a human tissue based on a CT-image. The developed probabilistic model is based on mixture density networks. It is empirically demonstrated that a convolutional U-net can learn a satisfactory mixture distribution in each pixel. To the best of our knowledge, this is the first implementation of mixture density networks working on images with a convolutional architecture. In Paper B, we construct and implement a new method called REMEDI for entropy estimation of continuous distributions using deep neural networks. Further, using empirical process theory, it is shown that such an estimator has a consistency property, ensuring that it has the theoretical capacity to estimate the entropy to arbitrary precision. The method is based on the celebrated Donsker-Varadhan lemma, a well-known fact from the theory of large deviations. The applicability of the method is demonstrated on distributions in moderate dimension, as well as on the task of training model in the information bottleneck framework, with satisfactory performance. In Paper C, we explore the integration of non-parametric model components into the flow-matching framework. This is done by learning a heavily compressed latent representation of images in the training dataset, which are then used as conditioning variables for the vector field network. Effectively, these can be seen as synthetic, continuous labels. The gain is a more efficient learning process, compared to baseline models, and more interpretable sampling. It is demonstrated that with sufficient compression, overfitting can be avoided, and diversity among samples attained, despite conditioning on training samples. Arguably, the most impactful area in machine learning is generative modeling. One part of the thesis deals with applying the theory of large deviations to two of its major methods, generative adversarial networks (GAN), and diffusion models, in particular Schrödinger bridges. The goal here is to prove large deviation principles, for certain sequences of probability measures associated with the models. In both cases, this allows a strong characterization of the convergence of these models, under varying certain model parameters, toward an idealized description of their behavior, often carrying a well-understood mathematical structure. In Schrödinger bridges, this idealized limit model consists of a dynamical optimal transport plan. This tells us that when varying the parameter in question, which is the reference noise level, or, as is often equivalent, the level of entropic regularization, the plans converge rapidly toward optimal transport behavior, justifying the interpretation of weakly regularized Schrödinger bridges as approximate optimal transport plans. Since Schrödinger bridges (or entropically regularized optimal transport plans) have nicer computational properties than optimal transport, they are often used in its place, and it is therefore important to understand how close this connection is. The large deviation principles derived here, applicable to several popular deep generative models, thus contribute to this understanding. Stating and proving such large deviations results are the contents of Papers E and F. For GANs, we consider recently developed particle systems making up cohorts of networks for the generative task, in Paper D.",
            "source_url": "https:\/\/openalex.org\/W7112536257",
            "pdf_url": "https:\/\/kth.diva-portal.org\/smash\/get\/diva2:2015506\/FULLTEXT02",
            "doi": "",
            "handle": "",
            "repository_name": "KTH Publication Database DiVA (KTH Royal Institute of Technology)",
            "source_record_id": "W7112536257",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 599,
            "title": "Generative modeling for ergodic systems",
            "title_ar": "",
            "title_en": "Generative modeling for ergodic systems",
            "author": "Claudia Drygala",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Germany",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "cnn",
            "categories": "deep-learning; generative-ai; cnn; gan; diffusion",
            "keywords": "",
            "abstract": "Chaotic dynamical systems are characterized by strong nonlinearity, aperiodic long-term behavior, and high sensitivity to initial conditions, which make their numerical simulation extremely demanding, especially in engineering-relevant applications such as turbulent flows. Although ergodic theory provides a statistical framework for describing such systems, resolving their dynamics through fine-grid simulations of ODEs and PDEs remains computationally expensive, while reduced-order closures often introduce substantial modeling error. This thesis investigates deep generative learning as an alternative approach for the synthetic modeling of invariant state snapshots of chaotic systems, with a particular focus on generative adversarial networks (GAN). By combining principles from ergodic theory with generative learning, this work establishes for the first time that GAN training converges mathematically for ergodic systems in the limit of long observation times. This result provides a rigorous theoretical foundation for the use of GAN in modeling chaotic systems. Building on this foundation, the thesis further shows that statistics computed from GAN-generated snapshots converge to those obtained from simulations, uniformly over all uniformly bounded functions. Together, these findings establish a solid framework for evaluating generative models of chaotic dynamics through physics-based statistical measures rather than explicitly embedding such measures into the training loss. The proposed methodology is applied to a hierarchy of systems of increasing complexity, including the Lorenz attractor, the double pendulum, the Kuramoto–Sivashinsky equation, and turbulent flow configurations such as flow around a cylinder and flow under periodic wake impact. Across these cases, deep convolutional GAN are shown to reproduce snapshots of chaotic and turbulent states with high fidelity, using relatively small training datasets and without incorporating physical boundary conditions. At the same time, they offer substantial computational advantages, generating new snapshots within seconds compared with the days often required for Large Eddy Simulations. The thesis also demonstrates that conditional GAN enable controlled generation and strong generalization across unseen flow configurations, even when the available training data is heavily reduced. In addition, it presents a comparative study of GAN, variational autoencoders, and denoising diffusion probabilistic models on both simulated LES data and experimental PIV data. Among these models, GAN provide the best overall trade-off between accuracy and efficiency, while diffusion models show strong potential for future data-rich applications. Finally, the work takes a first step beyond invariant snapshots toward spatio-temporal generative modeling of turbulence, opening promising directions for future research on time-dependent chaotic flow dynamics.",
            "source_url": "https:\/\/openalex.org\/W7214885496",
            "pdf_url": "",
            "doi": "10.14279\/depositonce-26707",
            "handle": "",
            "repository_name": "DepositOnce",
            "source_record_id": "W7214885496",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 600,
            "title": "High-resolution synthesis across domains: a wavelet-driven approach to generative modeling",
            "title_ar": "",
            "title_en": "High-resolution synthesis across domains: a wavelet-driven approach to generative modeling",
            "author": "SIGILLO, LUIGI",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Sapienza University of Rome",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Italy",
            "governorate": "",
            "publication_year": 2025,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "deep-learning; computer-vision; generative-ai; medical-ai; ai-applications; gan; diffusion; medical-imaging; ai-industry",
            "keywords": "",
            "abstract": "Deep generative modeling is revolutionizing visual data synthesis, from creative industry applications to tools for scientific discovery in fields like medical imaging and remote sensing. In image synthesis, deep learning models like Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Denoising Diffusion Probabilistic Models (DDPMs) help create novel, realistic, and controllable visual content. In specialized applications, these models can augment scarce data for medical analysis or enhance imagery for remote sensing tasks. A crucial challenge in applying deep learning to image synthesis is overcoming the persistent trade-offs between resolution, fidelity, and computational efficiency. While generative models have advanced significantly, they often struggle with generalization across diverse domains and data modalities. Conventional methods tend to scale poorly to ultra-high resolutions (UHR), leading to artifacts like repeated structures or blurred textures. Furthermore, these models often apply uniform refinement processes across all spatial regions, disregarding local frequency variations and failing to optimally allocate supervision to areas of different visual complexity. Wavelet transforms, particularly Discrete Wavelet Transforms (DWT) and their hypercomplex extension, Quaternion Wavelet Transforms (QWT), have shown promising results in multi-scale signal analysis. These methods operate by decomposing images into a hierarchy of frequency sub-bands, capturing both global structure and fine-grained details. This formulation endows models with the ability to leverage sparse representations and dimensionality reduction. However, their potential to reshape feature representation, inform conditioning, and adapt training objectives in a model-agnostic fashion had not been comprehensively exploited. In this thesis, we explore this concept and exploit the wavelet-driven learning paradigm to overcome the aforementioned shortcomings of traditional generative models. We leverage multi-scale analysis to make models inherently aware of frequency and spatial information. We first design a structure-aware GAN, StawGAN, tailored for cross-domain infrared-to-RGB image translation. Building on this foundation, we develop specialized diffusion models for domain-specific tasks, including high-fidelity maritime image super-resolution and efficient EEG-to-image synthesis. Moving beyond these conventional generative approaches, we introduce a series of wavelet-driven architectures that explicitly incorporate multi-scale signal representations. Among them is QUAVE, a novel framework leveraging quaternion wavelet transforms (QWT) to enhance feature extraction and improve generalization in medical imaging. Expanding on these insights, we also pioneer several wavelet-based super-resolution models: a QWT-conditioned diffusion model, a metadata- and wavelet-aware architecture for satellite imagery, and a highly efficient hybrid framework, Wavelet Diffusion GAN, that combines the strengths of GANs and diffusion processes. Finally, we address a direct extension of these works, focusing on high-fidelity synthesis. We culminate our investigation with a Latent Wavelet Diffusion (LWD) framework, a general and lightweight solution that enables existing latent diffusion and flow matching models to achieve UHR (up to 4K) synthesis without architectural modifications or additional inference costs. Through extensive experiments across various generative tasks involving different domains and data modalities, we have thoroughly explored the wavelet-driven paradigm while addressing scenario-specific challenges in fidelity, efficiency, and generalization, advancing research in this field.",
            "source_url": "https:\/\/openalex.org\/W7113426803",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "IRIS Research product catalog (Sapienza University of Rome)",
            "source_record_id": "W7113426803",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 601,
            "title": "Generative and geometric deep learning for 3D genome organization and biomolecular interaction modeling",
            "title_ar": "",
            "title_en": "Generative and geometric deep learning for 3D genome organization and biomolecular interaction modeling",
            "author": "Yanli Wang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Missouri",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "llm",
            "categories": "machine-learning; deep-learning; nlp; generative-ai; medical-ai; predictive-analytics; optimization; gnn; llm; gan; diffusion",
            "keywords": "",
            "abstract": "The three-dimensional organization of the genome and the interactions between biomolecules are fundamental to understanding cellular function and disease. Recent advances in sequencing technologies and structural biology have generated unprecedented amounts of biological data, creating new opportunities to study these processes. However, these data are often sparse, noisy, and highly complex, making it difficult to accurately model biological systems using conventional computational methods. This dissertation explores how generative and geometric deep learning can address these challenges by learning meaningful biological representations from high-dimensional genomic and structural data. The first part of this dissertation focuses on modeling three-dimensional genome organization from single-cell Hi-C data. Because single-cell Hi-C experiments capture only a small fraction of chromatin interactions, computational methods are essential for recovering missing information and revealing genome architecture. I first developed a deep residual generative adversarial network to enhance sparse single-cell Hi-C contact maps by recovering biologically meaningful chromatin interactions. Building on this work, I proposed HiCDiff, a diffusion model that iteratively denoises chromatin contact maps and produces more accurate representations of genome organization. These improved contact maps provide a stronger foundation for downstream analyses and structural modeling. To further characterize genome architecture, I developed an SO(3)-equivariant graph neural network that reconstructs three-dimensional chromosome structures directly from single-cell Hi-C data. By incorporating geometric equivariance into the learning process, the model captures the spatial relationships among genomic loci and reconstructs chromosome conformations with improved accuracy and biological consistency. The second part of this dissertation extends geometric deep learning to biomolecular interaction modeling. I developed MPBind, a multitask framework for predicting protein binding sites by integrating protein language model embeddings with equivariant graph neural networks. This approach combines evolutionary information encoded in protein sequences with three-dimensional structural features, enabling accurate prediction of diverse protein interactions while achieving strong generalization across previously unseen proteins. Although these studies address different biological problems, they share a common theme: combining generative models with geometry-aware neural networks to learn from complex biological data. Generative deep learning improves the quality of sparse and noisy experimental measurements, while geometric deep learning enables models to capture the structural principles underlying biological systems. Together, the methods presented in this dissertation provide a unified framework for studying genome organization and biomolecular interactions, demonstrating how modern artificial intelligence can advance computational genomics and structural biology and support future discoveries in biology and precision medicine.",
            "source_url": "https:\/\/openalex.org\/W7224088771",
            "pdf_url": "",
            "doi": "10.32469\/10355\/113055",
            "handle": "",
            "repository_name": "MOspace Institutional Repository (University of Missouri)",
            "source_record_id": "W7224088771",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 602,
            "title": "Music Generation with Generative Adversarial Networks",
            "title_ar": "",
            "title_en": "Music Generation with Generative Adversarial Networks",
            "author": "Li, Yupeng ; Linberg, Jonatan",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Jönköping University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Sweden",
            "governorate": "",
            "publication_year": 2023,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "deep-learning; generative-ai; gan",
            "keywords": "",
            "abstract": "At present, state-of-the-art deep learning music generation systems require a lot time and hardware resources to develop. This means that they are almost exclusively available to large companies. In order to reduce these requirements, more efficient techniques and methods need to be utilised. This project aims to investigate various approaches by developing a music generation system using generative adversarial networks, comparing different techniques and their effect on the system's performance. Our results show the difficulties in generating music in a more resource-constrained environment. We find that structuring the input space with conditional model constraints improves the systems' ability to conform to musical standards. The results also indicate the importance of a patch-based discriminator for evaluating the texture of the generated music. Finally, we propose a similarity loss as a way of reducing mode collapse in the generator, thus stabilising the training process.",
            "source_url": "https:\/\/openalex.org\/W7112384551",
            "pdf_url": "https:\/\/hj.diva-portal.org\/smash\/get\/diva2:1787759\/FULLTEXT01",
            "doi": "",
            "handle": "",
            "repository_name": "Jönköping University Publications (Jönköping University)",
            "source_record_id": "W7112384551",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 603,
            "title": "Generative Adversarial Networks for Classic Cryptanalysis",
            "title_ar": "",
            "title_en": "Generative Adversarial Networks for Classic Cryptanalysis",
            "author": "Deanne Charan",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "San Jose State University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2021,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; deep-learning; generative-ai; gan",
            "keywords": "",
            "abstract": "The necessity of protecting critical information has been understood for millennia. Although classic ciphers have inherent weaknesses in comparison to modern ciphers, many classic ciphers are extremely challenging to break in practice. Machine learning techniques, such as hidden Markov models (HMM), have recently been applied with success to various classic cryptanalysis problems. In this research, we consider the effectiveness of the deep learning technique CipherGAN---which is based on the well- established generative adversarial network (GAN) architecture---for classic cipher cryptanalysis. We experiment extensively with CipherGAN on a number of classic ciphers, and we compare our results to those obtained using HMMs.",
            "source_url": "https:\/\/openalex.org\/W4206278187",
            "pdf_url": "https:\/\/scholarworks.sjsu.edu\/cgi\/viewcontent.cgi?article=2034&context=etd_projects",
            "doi": "10.31979\/etd.h2mh-uh52",
            "handle": "",
            "repository_name": "San José State University ScholarWorks (San Jose State University)",
            "source_record_id": "W4206278187",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 604,
            "title": "Bildgenerierung mittels Generative Adversarial Networks",
            "title_ar": "",
            "title_en": "",
            "author": "Meline A. Ayvazyan",
            "supervisor": "",
            "degree_type": "master",
            "degree_evidence": "نص السجل: «Masterarbeit»",
            "university": "Saxon State and University Library Dresden",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Germany",
            "governorate": "",
            "publication_year": 2023,
            "language": "de",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "generative-ai; gan",
            "keywords": "",
            "abstract": "Das Hauptziel der Masterarbeit ist es, die Funktionsweise und das Verhalten von GANs in der Bildgenerierung zu untersuchen und dies anhand eines praktischen Beispiels zu verdeutlichen. Die Arbeit soll einen aktuellen Forschungsüberblick über die generative Modellierung, insbesondere im Bereich der Bildgenerierung, liefern und gleichzeitig die fachlichen Inhalte im Hinblick auf die Funktionsweise von GANs umfassend behandeln. Dabei werden sowohl theoretische Grundlagen als auch die praktische Umsetzung von GANs beleuchtet, um ein ganzheitliches Verständnis für dieses spannende Forschungsfeld zu vermitteln.",
            "source_url": "https:\/\/openalex.org\/W7139216419",
            "pdf_url": "https:\/\/tubaf.qucosa.de\/api\/qucosa%3A103059\/attachment\/ATT-0\/",
            "doi": "",
            "handle": "",
            "repository_name": "Qucosa (Saxon State and University Library Dresden)",
            "source_record_id": "W7139216419",
            "other_sources": "",
            "verification_status": "verified",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 605,
            "title": "Synthetic Ultrasound Video Generation with Generative Adversarial Networks",
            "title_ar": "",
            "title_en": "Synthetic Ultrasound Video Generation with Generative Adversarial Networks",
            "author": "Daniil Kulik",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Canada",
            "governorate": "",
            "publication_year": 2023,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; cnn; video-analysis; gan; medical-imaging",
            "keywords": "",
            "abstract": "Ultrasound is a non-invasive, radiation-free portable imaging modality that offers real-time diagnostics in different clinical settings.Ultrasound video analysis can benefit from the rise of AI-powered applications in healthcare.However, access to ultrasound data remains the main challenge for the development of state-of-the-art machine learning models.Synthetic data generation can provide various benefits to ultrasound imaging analysis.Namely, generating ultrasound videos with specific characteristics allows for better training and testing of machine learning models.This work proposes a generative adversarial network for conditional ultrasound video generation.We conduct a thorough quantitative and qualitative evaluation of the network.Additionally, we show the added value of using the synthetic video for data augmentation in a downstream task.Extensive experiments on the EchoNet-Dynamic dataset demonstrate that the proposed model achieves an FID score of 126 and an FVD score of 233 and can be used in data augmentation tasks in small data scenarios.Contents 7 Conclusion and Future Work 70 vi List of Tables 3.1 Prior work in ultrasound image and video generation with GANs. .5.1 Configuration of the image discriminator. . . . . . . . . . . . . . . .5.2 Configuration of the video discriminator. . . . . . . . . . . . . . . .5.3 Configuration of the transposed convolutional generator. . . . . . .5.4 Configuration of the subpixel generator. . . . . . .",
            "source_url": "https:\/\/openalex.org\/W4389205759",
            "pdf_url": "https:\/\/repository.library.carleton.ca\/downloads\/w0892c38h",
            "doi": "10.22215\/etd\/2023-15790",
            "handle": "",
            "repository_name": "Carleton University's Institutional Repository (MacOdrum Library, Carleton University)",
            "source_record_id": "W4389205759",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 606,
            "title": "Privacy Attacks and Protection in Generative Models",
            "title_ar": "",
            "title_en": "Privacy Attacks and Protection in Generative Models",
            "author": "HU, Hailong",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Luxembourg",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Luxembourg",
            "governorate": "",
            "publication_year": 2023,
            "language": "en",
            "research_domain": "generative-ai",
            "research_subdomain": "gan",
            "categories": "generative-ai; responsible-ai; ai-applications; gan; diffusion; fairness; ai-cybersecurity",
            "keywords": "",
            "abstract": "Recent years have witnessed the tremendous success of generative models in data synthesis. Typically, a well-trained model itself and its training set constitute key assets for model owners, which allows technology companies to gain a leading position in the global market. However, privacy is a key consideration in deploying state-of-the-art generative models in practice. On the one hand, the exposure of model privacy can lead to the compromise of the intellectual property rights of legitimate model owners, which consequently affects the market share of companies. On the other hand, the disclosure of training data, especially when it includes personal information, constitutes a direct infringement of data privacy, which severely leads to legal sanctions for companies. Indeed, the advent of emerging generative models critically necessitates novel privacy analysis and protection techniques to ensure the confidentiality of cutting-edge models and their training data. To solve these challenges, this dissertation investigates several new privacy attacks and protection methods for generative models from the perspective of model privacy and data privacy. In addition, this dissertation also explores a new mode that leverages existing pre-trained generative models to study the security vulnerabilities of discriminative models, which provides a fresh angle to apply generative models to the risk analysis of discriminative models. This dissertation is organized into three parts. In the first part, i.e. model privacy in generative models, I develop new model extraction attacks to steal generative adversarial networks (GANs). The evaluations show that preventing model extraction attacks against GANs is difficult but protecting GANs through verifying the ownership can be a deterrence against malicious adversaries. Thus, I further propose an ownership protection method to safeguard GANs, which can effectively recognize these stolen models constructed from physical stealing and model extraction. In the second part, i.e. data privacy in generative models, I develop two types of membership inference attacks against diffusion models, and the proposed loss-based method reveals the relationship between membership inference risks and the generative mechanism of diffusion models. I also investigate property inference risks in diffusion models and propose the first property aware sampling method to mitigate this attack, which bears the benefits of being plug-in and model-agnostic. In the third part, i.e. applications of generative models, I propose a new type of out-of-distribution (OOD) attack by leveraging off-the-shelf pre-trained GANs, which demonstrates that GANs can be utilized to directly construct samples to fool classification models and evade OOD detection. Taken together, this dissertation primarily provides new privacy attacks and protection methods for generative models and can contribute to a deeper and more comprehensive understanding of the privacy of generative artificial intelligence.",
            "source_url": "https:\/\/openalex.org\/W7110694972",
            "pdf_url": "https:\/\/orbilu.uni.lu\/bitstream\/10993\/58928\/1\/Dissertation_HailongHU.pdf",
            "doi": "",
            "handle": "",
            "repository_name": "Open Repository and Bibliography (University of Luxembourg)",
            "source_record_id": "W7110694972",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 607,
            "title": "Brain tumor segmentation via generative adversarial network",
            "title_ar": "",
            "title_en": "Brain tumor segmentation via generative adversarial network",
            "author": "Kevin Xie",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Illinois at Urbana-Champaign",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; computer-vision; generative-ai; medical-ai; predictive-analytics; image-segmentation; gan; medical-imaging",
            "keywords": "",
            "abstract": "With the development in technology and the complication of tumors presented in human brains, magnetic resonance imaging (MRI) has been an essential tool that physicians rely on to look at the structure of brain tissue. The ability to distinguish between normal tissue and tumor tissue has become the key to effective medical diagnosis and treatments. This thesis discusses developing machine learning techniques that will effectively segment the different parts of an MR image, namely the whole tumor, the tumor core and the enhancing tumor core. Using the dataset provided by the Perelman School of Medicine’s Center for Biomedical Image Computing & Analytics from University of Pennsylvania, a generative adversarial network (GAN) takes in imperfect predictions from other tumor segmentation networks and outputs more accurate and realistic predictions that resemble the ground truth. After our analysis, we see that the improvements that resulted from using a Pix2Pix GAN are shown in some cases, but the effects are unstable. Therefore, we conclude that fundamental changes in our GAN network are necessary to achieve more accurate results, and further research should be done on exploring the effects of using a Vox2Vox GAN.",
            "source_url": "https:\/\/openalex.org\/W3129104999",
            "pdf_url": "https:\/\/www.ideals.illinois.edu\/items\/132218\/bitstreams\/439091\/data.pdf",
            "doi": "",
            "handle": "",
            "repository_name": "Illinois Digital Environment for Access to Learning and Scholarship (University of Illinois at Urbana-Champaign)",
            "source_record_id": "W3129104999",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 608,
            "title": "Generative Adversarial Networks Based Reconstruction and Restoration of Cultural Heritage",
            "title_ar": "",
            "title_en": "Generative Adversarial Networks Based Reconstruction and Restoration of Cultural Heritage",
            "author": "Nesreen Hamadallah Jboor",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Qatar University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Qatar",
            "governorate": "",
            "publication_year": 2019,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "gan",
            "categories": "machine-learning; computer-vision; generative-ai; data-mining; gan",
            "keywords": "",
            "abstract": "Cultural heritage takes an important part in defining the identity and the history of a civilization or a nation. Valuing and preserving this heritage is thus a top priority for governments and heritage institutions. Through this paper, we present an image completion (inpainting) approach adapted for the curation and the completion of damaged artwork. Our approach uses a set of machine learning techniques such as Generative Adversarial Networks which are among the most powerful generative models that can be trained to generate realistic data samples. As we are focusing mostly on visual cultural heritage, the pipeline of our framework has many optimizations such as the use of clustering to optimize the training of the generative part to ensure a better performance across a variety of cultural data categories. The experimental results of our framework were validated on cultural dataset of paintings collected from Wiki-Art and the Rijksmuseum. We used the divide-and-conquer strategy by clustering the training data into different small clusters containing similarly looking images to train smaller Specialized DCGANs. The training has been made on five painting categories containing 2000 paintings each, which took an average of 6.1 training hours. Training the Specialized DCGAN on 1200 paintings from one of the clusters took 3.4 training hours. The inpainting results of the Specialized DCGANs are clearly better in quality than the results of a DCGAN trained on mixture of paintings or on painting category.",
            "source_url": "https:\/\/openalex.org\/W3004698630",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Qatar University QSpace (Qatar University)",
            "source_record_id": "W3004698630",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 609,
            "title": "Relaxed Wasserstein, Generative Adversarial Networks, Variational Autoencoders and their applications",
            "title_ar": "",
            "title_en": "Relaxed Wasserstein, Generative Adversarial Networks, Variational Autoencoders and their applications",
            "author": "Nan Yang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2019,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "gan",
            "categories": "deep-learning; computer-vision; generative-ai; gan",
            "keywords": "",
            "abstract": "Statistical divergences play an important role in many data-driven applications. Two notable examples are Distributionally Robust Optimization (DRO) problems and Generative Adversarial Networks (GANs).In the first section of my dissertation, we propose a novel class of statistical divergence called Relaxed Wasserstein (RW) divergence, which combines Wasserstein distance and Bregman divergence. We begin with its strong probabilistic properties, and then to illustrate its uses, we introduce Relaxed Wasserstein GANs (RWGANs) and compare it empirically with several state-of-the-art GANs in image generation. We show that it strikes a balance between training speed and image quality. We also discuss the potential use of Relaxed Wasserstein to construct ambiguity sets in DRO problems.In the second section of my dissertation, we show the application of another type of generative neural network, the Variational AutoEncoder (VAE), to metagenomic binning problems in bioinformatics. Shotgun sequencing is used to produce short reads from DNA sequences in a sample from a microbial community, which could contain thousands species of discovered or unknown microbes. The short reads are then assembled by connecting overlapping subsequences and thus forming longer sequences called contigs. Metagenomic binning is the process of grouping contigs from multiple organisms based on their genomes of origin. We propose a new network structure called MetaAE, which combines compositional and reference-based information in a nonlinear way. We show that this binning algorithm improves the performance of state-of-the-art binners by 20\\\\% on two independent synthetic datasets.",
            "source_url": "https:\/\/openalex.org\/W3205833864",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "eScholarship (California Digital Library)",
            "source_record_id": "W3205833864",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 610,
            "title": "Improve the efficiency of conditional generative models",
            "title_ar": "",
            "title_en": "Improve the efficiency of conditional generative models",
            "author": "Yanwu Xu",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Boston University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2024,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; computer-vision; generative-ai; medical-ai; transfer-learning; video-analysis; gan; diffusion; medical-imaging",
            "keywords": "",
            "abstract": "Deep generative models have undergone significant advancements, enabling the production of high-fidelity data across various fields, including computer vision and medical imaging. The availability of paired annotations facilitates a controllable generative process through conditional generative models, which capture the conditional distribution P(X|Y), where X represents high-dimensional data and Y denotes the associated annotation. This controllability makes conditional generative models more preferable than ordinary generative models, which only model P(X). For instance, the latest generative AI techniques within the Artificial Intelligence Generated Content (AIGC) realm have unlocked the potential for flexible image and video generation\/editing based on text descriptions or “prompts.” Additionally, generative AI has enhanced model efficiency by supplementing datasets with synthesized data in scenarios where annotations are unavailable or imprecise. Despite these capabilities, challenges persist in ensuring efficient training of conditional generative models. These challenges include: 1) capturing the intricate relationship between data and its annotations, which can introduce biases; 2) higher computational resource requirements during training and inference compared to discriminative models; and 3) the need to balance speed and quality at inference time. To address these challenges, this thesis introduces four models aimed at enhancing the training and inference efficiency of conditional generative models without compromising quality. The first method focuses on conditional Generative Adversarial Networks (cGANs), proposing a novel training objective to improve stability and diversity in synthetic data generation. The second method involves a hybrid generative model that combines GANs and Diffusion-based models to alleviate the unstable training of GANs and accelerate the denoising process. The third model introduces a fine-tuning framework that utilizes pre-trained diffusion parameters for high-fidelity, fast sampling, and quick adaptation during training. The final method presents a super-efficient 3D diffusion model for high-fidelity 3D CT synthesis, addressing the efficiency and quality gap in current models. These methods collectively tackle the efficiency of generative models and enhance generative quality in both computer vision and medical domains, suggesting sustainable solutions for the future of generative AI.",
            "source_url": "https:\/\/openalex.org\/W7126296160",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "OpenBU (Boston University)",
            "source_record_id": "W7126296160",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 611,
            "title": "Quality and Diversity in Generative Models through the lens of f-divergences",
            "title_ar": "",
            "title_en": "",
            "author": "Alexandre Vérine",
            "supervisor": "",
            "degree_type": "phd",
            "degree_evidence": "theses.fr: الفهرس الوطني الفرنسي لأطاريح الدكتوراه",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "France",
            "governorate": "",
            "publication_year": 2024,
            "language": "fr",
            "research_domain": "computer-vision",
            "research_subdomain": "gan",
            "categories": "machine-learning; computer-vision; generative-ai; gan",
            "keywords": "",
            "abstract": "Explorer la qualité et la diversité des modèles génératifs Les modèles génératifs sont devenus un outil essentiel dans l'apprentissage automatique pour générer des échantillons réalistes à partir de distributions de données complexes. Malgré des avancées significatives dans les modèles tels que les Generative Adversarial Network, les Variational Autoencoders, les Normalizing Flows et les modèles de diffusion, des défis persistent pour régler le compromis entre la qualité et la diversité des échantillons. Cette thèse aborde le problème fondamental de la caractérisation, l'ajustement et de l'amélioration de la qualité et de la diversité dans les modèles génératifs. La précision et le rappel ont émergé comme des métriques cruciales pour évaluer la qualité et la diversité des modèles génératifs. La précision mesure combien d'échantillons générés sont réalistes avec la distribution de données réelle, reflétant la qualité des échantillons. Le rappel évalue combien d'échantillons de la distribution de données réelle peuvent être générés, indiquant la diversité des échantillons. La première contribution majeure de ce travail est l'unification des définitions de la précision et du rappel dans le cadre des f-divergences. En exprimant les métriques les plus populaires et leurs dérivés en tant qu'un famille de f-divergence, la PR-Divergence, nous établissons un système d'évaluation cohérent et complet pour les modèles génératifs. Cette formulation théorique permet une compréhension plus claire et une mesure plus précise des performances des modèles en termes de qualité et de diversité. En s'appuyant sur cette base théorique, la thèse introduit une méthode novatrice pour estimer la PR-divergence de manière différentiables, facilitant son utilisation comme fonction objective dans la formation des modèles génératifs. Cette approche permet d'optimiser n'importe quel compromis spécifique entre précision et rappel. Cette méthode se montre complémentaire aux méthodes existantes. De plus, la thèse propose une méthode optimale d'échantillonnage par rejet qui améliore à la fois la précision et le rappel. Cette méthode est démontrée comme étant optimale en termes de toute f-divergence, fournissant une technique robuste pour affiner les sorties des modèles génératifs pré entraînés. L'algorithme d'échantillonnage par rejet est conçu pour fonctionner sous des budgets computationnels limités, le rendant pratique pour des applications réelles .La validation expérimentale des méthodes proposées est réalisée sur une variété de jeux de données, incluant MNIST, CIFAR-10, Fashion MNIST, CelebA, FFHQ et ImageNet. En utilisant les Normalizing Flows, les Generative Adversarial Networks et les modèles de diffusion, nous démontrons l'efficacité de nos approches pour ajuster le compromis entre la qualité et la diversité des échantillons générés, puis pour améliorer la qualité. Les résultats soulignent la supériorité de nos méthodes par rapport aux métriques traditionnelles et aux techniques existantes.",
            "source_url": "https:\/\/openalex.org\/W7151888075",
            "pdf_url": "",
            "doi": "10.70675\/36e5bb7cz8351z4ba7z8fffzcc9523acc864",
            "handle": "",
            "repository_name": "theses.fr (ABES)",
            "source_record_id": "W7151888075",
            "other_sources": "",
            "verification_status": "verified",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 612,
            "title": "Deformable Image Registration Using Attentional Generative Adversarial Networks",
            "title_ar": "",
            "title_en": "Deformable Image Registration Using Attentional Generative Adversarial Networks",
            "author": "Hanchong Zhou",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "",
            "governorate": "",
            "publication_year": 2021,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; predictive-analytics; cnn; gan; medical-imaging",
            "keywords": "",
            "abstract": "Deformable image registration is a fundamental process that aims to estimate non-linear spatial correspondence between input images. Medical image registration serves wildly on clinical treatment evaluation, monitoring disease and tracking disease. Conventional registration algorithms iteratively optimize a similarity function for each pair of images, which result in long registration time. Learning based registration method typically use convolutional neural networks to learn features automatically during training and register an image pair in one shot. Generative adversarial network is a novel structure that involves a generator and a discriminator, where the discriminator encourages the former to generate better results. Predicting deformation between brain magnetic resonance images is a complicated task because of their high-dimensional non-linear transform. A generative model leveraging is proposed using an attentional mechanism to estimate the complicated deformation field, and train the model using perceptual cyclic constraints. As an unsupervised method, our model dose not need any labels for training. Experimental results show quantitative evidence that the proposed method can predict reliable deformation field at a fast speed.",
            "source_url": "https:\/\/openalex.org\/W3124900387",
            "pdf_url": "https:\/\/ucalgary.scholaris.ca\/bitstreams\/de143ab7-7823-45a9-bfac-1a2793663276\/download",
            "doi": "10.11575\/prism\/38593",
            "handle": "",
            "repository_name": "",
            "source_record_id": "W3124900387",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 613,
            "title": "Dual-Attention Generative Adversarial Network and Flame and Smoke Analysis",
            "title_ar": "",
            "title_en": "Dual-Attention Generative Adversarial Network and Flame and Smoke Analysis",
            "author": "Yuchuan Li",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Ottawa",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Canada",
            "governorate": "",
            "publication_year": 2021,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "transformers",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; expert-systems; predictive-analytics; transformers; image-segmentation; video-analysis; gan",
            "keywords": "",
            "abstract": "Flame and smoke image processing and analysis could improve performance to detect smoke or fire and identify many complicated fire hazards, eventually to help firefighters to fight fires safely. Deep Learning applied to image processing has been prevailing in recent years among image-related research fields. Fire safety researchers also brought it into their studies due to its leading performance in image-related tasks and statistical analysis. From the perspective of input data type, traditional fire research is based on simple mathematical regressions or empirical correlations relying on sensor data, such as temperature. However, data from advanced vision devices or sensors can be analyzed by applying deep learning beyond auxiliary methods in data processing and analysis. Deep Learning has a bigger capacity in non-linear problems, especially in high-dimensional spaces, such as flame and smoke image processing. We propose a video-based real-time smoke and flame analysis system with deep learning networks and fire safety knowledge. It takes videos of fire as input and produces analysis and prediction for flashover of fire. Our system consists of four modules. The Color2IR Conversion module is made by deep neural networks to convert RGB video frames into InfraRed (IR) frames, which could provide important thermal information of fire. Thermal information is critically important for fire hazard detection. For example, 600 °C marks the start of a flashover. As RGB cameras cannot capture thermal information, we propose an image conversion module from RGB to IR images. The core of this conversion is a new network that we innovatively proposed: Dual-Attention Generative Adversarial Network (DAGAN), and it is trained using a pair of RGB and IR images. Next, Video Semantic Segmentation Module helps extract flame and smoke areas from the scene in the RGB video frames. We innovated to use synthetic RGB video data generated and captured from 3D modeling software for data augmentation. After that, a Video Prediction Module takes the RGB video frames and IR frames as input and produces predictions of the subsequent frames of their scenes. Finally, a Fire Knowledge Analysis Module predicts if flashover is coming or not, based on fire knowledge criteria such as thermal information extracted from IR images, temperature increase rate, the flashover occurrence temperature, and increase rate of lowest temperature. For our contributions and innovations, we introduce a novel network, DAGAN, by applying foreground and background attention mechanisms in the image conversion module to help reduce the hardware device requirement for flashover prediction. Besides, we also make use of combination of thermal information from IR images and segmentation information from RGB images in our system for flame and smoke analysis. We also apply a hybrid design of deep neural networks and a knowledge-based system to achieve high accuracy. Moreover, data augmentation is also applied on the Video Semantic Segmentation Module by introducing synthetic video data for training. The test results of flashover prediction show that our system has leading places quantitative and qualitative in terms of various metrics compared with other existing approaches. It can give a flashover prediction as early as 51 seconds with 94.5% accuracy before it happens.",
            "source_url": "https:\/\/openalex.org\/W3204711149",
            "pdf_url": "",
            "doi": "10.20381\/ruor-26991",
            "handle": "",
            "repository_name": "uO Research (University of Ottawa)",
            "source_record_id": "W3204711149",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 667,
            "title": "Video-based similar gesture action recognition using deep learning and GAN-based approaches",
            "title_ar": "",
            "title_en": "Video-based similar gesture action recognition using deep learning and GAN-based approaches",
            "author": "Di Wu",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Technology Sydney",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Australia",
            "governorate": "",
            "publication_year": 2019,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "video-analysis",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; robotics; responsible-ai; predictive-analytics; cnn; rnn; video-analysis; gan; fairness",
            "keywords": "",
            "abstract": "Human action is not merely a matter of presenting patterns of motion of different parts of the body, in addition, it is also a description of intention, emotion and thoughts of the person. Hence, it has become a crucial component in human behavior analysis and understanding. Human action recognition has a wide variety of applications such as surveillance, robotics, health care, video searching and human-computer interaction. Analysing human actions manually is tedious and easily prone to errors. Therefore, computer scientists have been trying to bring the abilities of cognitive video understanding to human action recognition systems by using computer vision techniques. However, human action recognition is a complex task in computer vision because of the camera motion, occlusion, background cluttering, viewpoint variation, execution rate and similar gestures. These challenges significantly degrade the performance of the human action recognition system. The purpose of this research is to propose solutions based on traditional machine learning methods as well as the state-of-the-art deep learning methods to automatically process video-based human action recognition. This thesis investigates three research areas of video-based human action recognition: traditional human action recognition, similar gesture action recognition, and data augmentation for human action recognition. To start with, the feature-based methods using classic machine learning algorithms have been studied. Recently, deep convolutional neural networks (CNN) have taken their place in the computer vision and human action recognition research areas and have achieved tremendous success in comparison to traditional machine learning techniques. Current state-of-the-art deep convolutional neural networks were used for the human action recognition task. Furthermore, recurrent neural networks (RNN) and its variation of long-short term memory (LSTM) are used to process the time series features which are handcrafted features or extracted from the CNN. However, these methods suffer from similar gestures, which appear in the human action videos. Thus, a hierarchical classification framework is proposed for similar gesture action recognition, and the performance is improved by the multi-stage classification approach. Additionally, the framework has been modified into an end-to-end system, therefore, the similar gestures can be processed by the system automatically. In this study, a novel data augmentation framework for action recognition has been proposed, the objective is to generate well learnt video frames from action videos which can enlarge the dataset size as well as the feature bias. It is very important for a human action recognition system to recognize the actions with similar gestures as accurately as possible. For such a system, a generative adversarial net (GAN) is applied to learn the original video datasets and generate video frames by playing an adversarial game. Furthermore, a framework is developed for classifying the original dataset in the first place to obtain the confusion matrix using a CNN. The similar gesture actions will be paired based on the confusion matrix results. The final classification result will be applied on the fusion dataset which contains both original and generated video frames. This study will provide realtime and practical solutions for autonomous human action recognition system. The analysis of similar gesture actions will improve the performance of the existing CNN-based approaches. In addition, the GAN-based approaches from computer vision have been applied to the graph embedding area, because graph embedding is similar to image embedding but used for different purposes. Unlike the purpose of the GAN in computer vision for generating the images, the GAN in graph embedding can be used to regularize the embedding. So the proposed methods are able to reconstruct both structural characteristics and node features, which naturally possess the interaction between these two sources of information while learning the embedding.",
            "source_url": "https:\/\/openalex.org\/W3020614990",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "UTS ePRESS (University of Technology Sydney)",
            "source_record_id": "W3020614990",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:51",
            "updated_at": "2026-11-15 16:52:51"
        },
        {
            "id": 677,
            "title": "Transfer Learning Methods for Object and Action Recognition",
            "title_ar": "",
            "title_en": "Transfer Learning Methods for Object and Action Recognition",
            "author": "Samitha Herath",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Australian National University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Australia",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "video-analysis",
            "categories": "machine-learning; computer-vision; generative-ai; ai-applications; transfer-learning; few-zero-shot; video-analysis; gan; ai-education",
            "keywords": "",
            "abstract": "Thanks to the advancement of machine learning and computer vision research we are observing remarkable improvements in object and action recognition. However, it can be seen that when there are differences (e.g., distributional, representations, semantic) in the training and test data, many learning algorithms fail to perform well or at times becomes infeasible. Nevertheless, problem setups dealing specifically with such differences are constantly being explored in literature. In general these problems are known as transfer learning problems. For instance, the objective of Domain Adaptation (DA) is to learn models that are invariant to distributional differences between train and test data (i.e., source and target domain data). Zero-Shot Learning (ZSL) is a framework for recognizing instances from classes never seen during training (i.e., missing classes in training data). In ZSL, the idea is to use related semantic information about data to workaround this data deficiency. In this thesis we present two solutions for Domain Adaptation and Zero-Shot Learning. First, we discuss our DA solution for learning an invariant Hilbert space for source and target domain data embedding. This is achieved by learning projections from individual domains to a latent space. We learn a Mahalanobis metric on this latent space to simultaneously minimizing domain disparity while maximizing a measure of discriminative power. We also show that the alignment of statistical properties (e.g., first and second order statistics) helps in improving our solution's performance, especially when there are no labeled samples from the target domain. We explore the effect of geometrical constraints such as orthogonality and positive definiteness for our model. Here, we use Riemannian optimization methods to seamlessly implement these constraints in our model training. Inspired by the success of statistical alignment, we further explore statistical alignment by min-max learning. Our motivation here is to avoid the trivial feature degeneration occurring when aligning two spaces by minimizing a statistical loss such as KL divergence. Here, we show that min-max statistical alignment is a promising solution for DA as well as ZSL. In fact, we show that min-max statistical alignment can reach comparable performance with state of the art adversarial solutions, including the ones using powerful Wasserstein GANs as the backbone. Lastly, we discuss our work on spatio-temporal transfer learning for human action recognition. In our survey on human action recognition from videos, we extensively study representation based and deep network based action recognition solutions. Here, we observe that one of the high-impact video action recognition solutions is the Two-Stream networks solution. However, the applicability of two-stream solution requires a temporal stream. In other words this restricts its applicability only to video data. In this work we explore use of transfer learning methods for devising two stream action recognition for still action images where there are no temporal information. We call this Temporal to Still Image Learning (or in short T2SIL). To explore the T2SIL framework we use existing still image and video datasets to form a new human action dataset of 40 classes. The data set contains still image actions and a video actions counterpart with the same set of classes. In this work, we explore three state of the art transfer learning solutions (i.e., Deep embedding learning, GANs and model distillation with Teacher-Student Networks) for T2SIL.",
            "source_url": "https:\/\/openalex.org\/W3036155538",
            "pdf_url": "",
            "doi": "10.25911\/5ee74e7c697e0",
            "handle": "",
            "repository_name": "ANU Open Research (Australian National University)",
            "source_record_id": "W3036155538",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 701,
            "title": "Mitigating Poisoning Threats and Their Privacy Implications in Federated Learning",
            "title_ar": "",
            "title_en": "Mitigating Poisoning Threats and Their Privacy Implications in Federated Learning",
            "author": "Chamara Sandeepa",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University College Dublin",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Ireland",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; generative-ai; medical-ai; responsible-ai; ai-applications; data-mining; gan; xai; federated-learning; fairness; ai-cybersecurity; ai-industry",
            "keywords": "",
            "abstract": "Advancements in Machine Learning (ML) and distributed intelligence have accelerated the adoption of Federated Learning (FL) as a key enabler of privacy-preserving applications in domains such as IoT, healthcare, and Beyond 5G networks. FL enables collaborative model training without centralising data, but its distributed and heterogeneous nature introduces new security and privacy challenges. In addition to conventional adversarial threats, FL faces unique vulnerabilities including poisoning attacks, and privacy leakage through model updates via threats like inference attacks. Addressing these challenges is critical for ensuring trust in next-generation distributed AI systems. This thesis investigates such threats in FL with a particular emphasis on poisoning attacks and the interplay between poisoning and privacy leakage. The work also investigates how poisoning attacks can be mitigated, essentially preventing any privacy vulnerabilities induced via poisoning. To this end, four major contributions are presented: Rec-Def, SHERPA, X-SHIELD, and ROAM. First, Rec-Def is proposed as a defence against GAN-based inference attacks, highlighting the connection between poisoning and privacy risks. Second, SHERPA introduces an interpretable aggregation algorithm that employs Explainable AI (XAI) technique called SHAP feature attributions to detect poisoned model updates, improving the transparency and reliability of poisoning detection. Third, X-SHIELD extends this approach to hierarchical FL, combining explanation-guided detection with scalable clustering to address robustness in large-scale FL deployments. Finally, the thesis explores the extension of defences and its limitations on fully Decentralised FL (DFL) architectures. For this, the thesis proposes an offensive use of XAI via an new stealthy poisoning attack called ROAM, based on Layerwise Relevance Propagation (LRP) and designs a countermeasure that perturbs sensitive neurons to mitigate privacy leakage via poisoning. The effectiveness of the proposed approaches is validated through extensive experiments across multiple datasets, learning scenarios, and adversarial settings. Results demonstrate that the solutions enhance robustness and privacy while remaining scalable and adaptable to diverse FL environments. Overall, this thesis contributes to advancing the safety and trustworthiness of FL by investigating novel vulnerabilities introducing and solutions that balance privacy preservation, poisoning robustness, and justifiability. These contributions provide both theoretical insights and design mechanisms for deploying FL securely in critical infrastructures such as 6G networks, IoT ecosystems, and industrial applications.",
            "source_url": "https:\/\/openalex.org\/W7165993495",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Research Repository UCD (University College Dublin)",
            "source_record_id": "W7165993495",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 703,
            "title": "Robust and Privacy-Preserving Federated Learning",
            "title_ar": "",
            "title_en": "",
            "author": "Fatima Elhattab",
            "supervisor": "",
            "degree_type": "phd",
            "degree_evidence": "theses.fr: الفهرس الوطني الفرنسي لأطاريح الدكتوراه",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "France",
            "governorate": "",
            "publication_year": 2023,
            "language": "fr",
            "research_domain": "medical-ai",
            "research_subdomain": "gan",
            "categories": "machine-learning; generative-ai; medical-ai; responsible-ai; ai-applications; gan; federated-learning; fairness; ai-cybersecurity; ai-industry",
            "keywords": "",
            "abstract": "Vers un Apprentissage fédéré robuste et préservant la confidentialité Dans le monde numérique en perpétuelle mutation d’aujourd’hui, l’apprentissage automatique est désormais une puissance essentielle et révolutionnaire, comme le démontrent de multiples recherches. Son impact profond s’étend à travers diverses industries, offrant des solutions et des innovations révolutionnaires qui ont remodelé la manière dont nous interagissons avec la technologie et prenons des décisions. Des systèmes de recommandation améliorant la diffusion de contenu sur les plateformes à la présence d’assistants personnels virtuels comme Siri et Alexa, capables de comprendre et de répondre à des commandes en langage naturel. Dans des domaines tels que la santé, il contribue au diagnostic des maladies, tandis que dans la finance, il renforce la détection de la fraude et l’évaluation des risques. Cette ubiquité de l’apprentissage automatique signifie non seulement une tendance technologique, mais aussi un changement fondamental dans les approches de résolution de problèmes et de prise de décisions. Cependant, cette vague d’innovation axée sur les données a soulevé une préoccupation primordiale : la protection de la vie privée des individus et de leurs données personnelles. Le Règlement général sur la protection des données (RGPD) illustre l’importance accrue de la protection des données à l’ère moderne. L’apprentissage fédéré représente un paradigme prometteur en apprentissage automatique, permettant la formation collaborative de modèles entre des appareils décentralisés. Cependant, il présente une vulnérabilité à diverses attaques. Cette recherche est divisée en deux axes principaux, chacun abordant des défis cruciaux en matière de sécurité et de confidentialité dans le contexte de l’apprentissage fédéré. Le premier axe se concentre sur la lutte contre les attaques d’empoisonnement pour un apprentissage fédéré robuste, où les adversaires cherchent à introduire des tâches nuisibles dans les modèles fédérés en plus de leurs tâches principales. Pour détecter ces attaques, on introduit ARMOR, un nouveau système de détection d’attaque basé sur GAN qui analyse les informations intégrées dans les mises à jour du modèle. Le deuxième axe concerne la lutte contre les attaques d’inférence pour l’apprentissage fédéré préservant la vie privée, en particulier les attaques d’inférence d’appartenance. Pour renforcer la confidentialité en apprentissage fédéré, deux approches novatrices sont introduites : PASTEL, qui améliore la résilience des systèmes d’apprentissage fédéré contre les MIAs en minimisant la différence de généralisation interne, et DINAR, une méthode d’apprentissage fédéré préservant la confidentialité à grain fin qui obscurcit les couches sensibles à la confidentialité et utilise une descente de gradient adaptative pour améliorer l’utilité du modèle. Ces objectifs de recherche visent collectivement à relever les défis en matière de sécurité et de confidentialité et à faire progresser le domaine de l'apprentissage fédéré.",
            "source_url": "https:\/\/openalex.org\/W7151959159",
            "pdf_url": "http:\/\/www.theses.fr\/2023ISAL0096\/document",
            "doi": "10.70675\/062161c5z9cdez4465z92bezb97b4a23cb36",
            "handle": "",
            "repository_name": "theses.fr (ABES)",
            "source_record_id": "W7151959159",
            "other_sources": "",
            "verification_status": "verified",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 738,
            "title": "AI-Generated Image Detection Using Multimodal Deep Learning",
            "title_ar": "",
            "title_en": "AI-Generated Image Detection Using Multimodal Deep Learning",
            "author": "Mamadou Keïta",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "France",
            "governorate": "",
            "publication_year": 2026,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "llm",
            "categories": "machine-learning; deep-learning; computer-vision; nlp; generative-ai; responsible-ai; few-zero-shot; transformers; multimodal; llm; gan; diffusion; xai; fairness",
            "keywords": "",
            "abstract": "Generative artificial intelligence (AI) and its applications offer significant benefits but raise potentially critical societal and ethical concerns. This thesis explores the field of synthetic content generation and detection, with a particular focus on synthetic images. These are AI-generated images that can appear indistinguishable from real photographs, creating new challenges for digital forensics, media integrity, and public trust. While generative models such as generative adversarial networks (GANs) and diffusion models have evolved rapidly, most existing detection methods remain limited in terms of generalizability, robustness, and interpretability. To address these challenges, this research investigates new AI-driven frameworks to improve the generalizability, robustness, and interpretability of current image detection methods.The various aspects of synthetic image detection are covered in this thesis by four key contributions. The first contribution, Bi-LORA, presents an efficient vision-language approach that reformulates the detection problem into an image caption task, achieving strong zero-shot generalization across unseen generative models. The second, RAVID, proposes a retrieval-augmented visual detection framework that enhances robustness and interpretability by incorporating external visual context. The third, DeeCLIP, presents a lightweight transformer-based model that fuses shallow and deep features to improve resistance to image degradation and post-processing. Finally, FIDAVL unifies synthetic image detection and source attribution in a single multitasking framework based on the soft prompt tuning of vision-language models.This thesis provides an in-depth analysis of the current state of synthetic image research and its societal impacts. It emphasizes the urgent need for effective and robust detection methods. The approaches introduced advance the scientific community and could benefit practical applications in media authenticity and security. This research marks an important step forward in developing solutions to the ethical challenges posed by GenAI.",
            "source_url": "https:\/\/openalex.org\/W7126004607",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "HAL",
            "source_record_id": "W7126004607",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 749,
            "title": "Adaptive Fusion Techniques for Effective Multimodal Deep Learning",
            "title_ar": "",
            "title_en": "Adaptive Fusion Techniques for Effective Multimodal Deep Learning",
            "author": "Gaurav Sahu",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "University of Waterloo",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Canada",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "deep-learning",
            "research_subdomain": "transformers",
            "categories": "deep-learning; computer-vision; nlp; generative-ai; transformers; multimodal; video-analysis; sentiment-analysis; machine-translation; speech-recognition; gan",
            "keywords": "",
            "abstract": "Effective fusion of data from multiple modalities, such as video, speech, and text, is a challenging task due to the heterogeneous nature of multimodal data. In this work, we propose fusion techniques that aim to model context from different modalities effectively. Instead of defining a deterministic fusion operation, such as concatenation, for the network, we let the network decide “how” to combine given multimodal features more effectively. We propose two networks: 1) Auto-Fusion network, which aims to compress information from different modalities while preserving the context, and 2) GAN-Fusion, which regularizes the learned latent space given context from complementing modalities. A quantitative evaluation on the tasks of multimodal machine translation and emotion recognition suggests that our adaptive networks can better model context from other modalities than all existing methods, many of which employ massive transformer-based networks.",
            "source_url": "https:\/\/openalex.org\/W3084985854",
            "pdf_url": "https:\/\/uwspace.uwaterloo.ca\/bitstreams\/d8ed4daf-ffb2-4594-8566-a7a3085f4b99\/download",
            "doi": "",
            "handle": "",
            "repository_name": "UWSpace (University of Waterloo)",
            "source_record_id": "W3084985854",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 770,
            "title": "Application of reliability and resilience models to machine learning: a thesis in Computer Engineering",
            "title_ar": "",
            "title_en": "Application of reliability and resilience models to machine learning: a thesis in Computer Engineering",
            "author": "Zakaria Faddi",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "",
            "governorate": "",
            "publication_year": 2023,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "image-classification",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; ai-applications; predictive-analytics; cnn; image-classification; gan; ai-cybersecurity",
            "keywords": "",
            "abstract": "Machine Learning (ML) systems such as Convolutional Neural Networks (CNNs) are susceptible to adversarial scenarios. In these scenarios, an attacker attempts to manipulate or deceive a machine learning model by providing malicious input resulting in incorrect predictions or decisions, which can have severe consequences in security, healthcare, and finance applications. Failures in the ML algorithm can lead to failures in the application domain and the system to which they provide functionality, which may possess performance requirements, hence the need for software reliability and resilience assessment. Many studies propose enhanced techniques to improve the robustness and resilience of ML algorithms. Yet, very few provide quantitative methods that promote risk assessment or measure progress toward improvements in such systems. This research demonstrates the applicability of software reliability and resilience tools to ML algorithms providing an objective approach to assess reliability and resilience recovery after a degradation from known adversarial attacks. An image classification model using CNNs has been created as the target of two generative adversarial attacks (i) the Fast Gradient Sign Method and (ii) the Projected Gradient Descent. The classifier and adversarial attacks were utilized in a min-max game scenario to replicate incremental learning and adaptive adversarial training, to minimize\/maximize the number of failures of the classification algorithm. Model-generated data was collected and prepared in a format suitable for software reliability growth models (SRGM) with and without covariates and resilience models, including failure counts, memory usage, and the intensity of noise added by the attacks. Our results suggest that tools incorporating SRGMs and resilience techniques are suitable for quantifying and predicting the resilience and reliability of ML models and may find practical use in many domains in which these models are applied.",
            "source_url": "https:\/\/openalex.org\/W4409527456",
            "pdf_url": "",
            "doi": "10.62791\/20291",
            "handle": "",
            "repository_name": "",
            "source_record_id": "W4409527456",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:52",
            "updated_at": "2026-11-15 16:52:52"
        },
        {
            "id": 836,
            "title": "Deep Learning in Medical Image Analysis: Alzheimer’s Disease Classification & Whole Fetal Brain Segmentation.",
            "title_ar": "",
            "title_en": "Deep Learning in Medical Image Analysis: Alzheimer’s Disease Classification & Whole Fetal Brain Segmentation.",
            "author": "Jesús Alejandro Valdés Valdés",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Tecnológico de Monterrey",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Mexico",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "medical-ai",
            "research_subdomain": "medical-imaging",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; medical-ai; transfer-learning; cnn; image-segmentation; gan; medical-imaging",
            "keywords": "",
            "abstract": "During the last couple years there has been a growing interest in Deep Learning (DL), thanks in part to an increasing availability of computational power and of massive labeled datasets. This work presents two different Deep Learning applications for medical image analysis related tasks. The first application is the classification of Alzheimer Disease (AD) using MRI. AD is the most common form of dementia. However, it’s difficult to diagnose. During this work three different DL models are proposed to classify subjects into either AD or Normal Control (NC) groups. The models proposed are a custom CNN model, vgg16 + Transfer Learning and vgg16 + Fine Tuning. The vgg16 based methods make use of pre-learned weights trained on the ImageNet dataset. All of the models used as an input a single 2D MRI slice. This helped reduced model complexity and training times. After performing a 10 times repeated 5-Fold cross validation it was discovered that the best model was vgg16 with Fine Tuning. It reached an accuracy of 80% and an AUC of 0.86. The obtained results proved that the use of weights learned from different domains could be useful in medical image applications. The main contributions of this section were the use of a single MRI slice to classify patients, the model validation technique and the use of transfer learning. The second application is the fully automatization of the Whole Fetal Brain Segmentation in maternal MRI scans. For this U-Net [41] and other segmentation networks were compared. The U-Net model was also modified to include two types of attention components. Squeeze and Excitation module [16] and Attention Gates [39]. To avoid overfitting the models were trained using rigorous regularization (channel wise dropout, weight decay and data augmentation). They were also trained using different loss functions, including a Hybrid (BCE+Dice) loss function. The results of the 10-fold cross validation experiments showed statistically that no model was able to outperform the vanilla U-Net (DSC of 0.93 and HD 7.59mm). The proposed Attention Gated U-Net trained with the Hybrid loss reached a 0.95 DSC and 5.03mm HD (no significant difference). However, results on challenging data show a better performance for the Attention Gated models. This led to the conclusion that the possible advantage of using attention gates outweighed the cost of adding the attention module. The results were also improved by the use of a post processing step based on traditional computer vision techniques. The main contributions to this problem were the use of 10-fold cross validation, the statistical comparison between models, the use of the Hausdorff distance metric, the post processing step and the addition of attention components. This work also provides an overview of other deep learning applications also used within a medical setting. The techniques mentioned include image labeling, autoencoders and GANs for data augmentation. It also provides an overview of the current challenges and limitations DL faces and the most common way of mitigating them. Thanks to the research conducted during this thesis it can be concluded that the use of DL can be used in a medical setting focused on the analysis of images. Particularly on classification and segmentation tasks. However, certain challenges such as having a large enough data set and availability of the required hardware to carry out research must be considered before attempting to solve a problem with DL.",
            "source_url": "https:\/\/openalex.org\/W3014475976",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "Repositorio del Tecnológico de Monterrey (Tecnológico de Monterrey)",
            "source_record_id": "W3014475976",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:53",
            "updated_at": "2026-11-15 16:52:53"
        },
        {
            "id": 1001,
            "title": "Deep Learning Based Face Detection and Recognition in MWIR and Visible Bands",
            "title_ar": "",
            "title_en": "Deep Learning Based Face Detection and Recognition in MWIR and Visible Bands",
            "author": "Suha Reddy Mokalla",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "West Virginia University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2020,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "face-recognition",
            "categories": "machine-learning; deep-learning; computer-vision; generative-ai; transfer-learning; face-recognition; gan",
            "keywords": "",
            "abstract": "In non-favorable conditions for visible imaging like extreme illumination or nighttime, there is a need to collect images in other spectra, specifically infrared. Mid-Wave infrared (3-5 microm) images can be collected without giving away the location of the sensor in varying illumination conditions. There are many algorithms for face detection, face alignment, face recognition etc. proposed in visible band till date, while the research using MWIR images is highly limited. Face detection is an important pre-processing step for face recognition, which in turn is an important biometric modality. This thesis works towards bridging the gap between MWIR and visible spectrum through three contributions. First, a dual band based deep face detection model that works well in visible and MWIR spectrum is proposed using transfer learning. Different models are trained and tested extensively using visible and MWIR images and the one model that works well for this data is determined. For this model, experiments are conducted to learn the speed\/accuracy trade-off. Following this, the available MWIR dataset is extended through augmentation using traditional methods and generative adversarial networks (GANs). Traditional methods used to augment the data are brightness adjustment, contrast enhancement, applying noise to and de-noising the images. A deep learning based GAN architecture is developed and is used to generate new face identities. The generated images are added to the original dataset and the face detection model developed earlier is once again trained and tested. The third contribution is the proposal of another GAN that converts given thermal ace images into their visible counterparts. A pre-trained model is used as discriminator for this purpose and is trained to classify the images as real and fake and an identity network is used to provide further feedback to the generator. The generated visible images are used as probe images and the original visible images are used as gallery images to perform face recognition experiments using a state-of-the-art visible-to-visible face recognition algorithm.",
            "source_url": "https:\/\/openalex.org\/W3024705230",
            "pdf_url": "https:\/\/researchrepository.wvu.edu\/cgi\/viewcontent.cgi?article=8577&context=etd",
            "doi": "10.33915\/etd.7632",
            "handle": "",
            "repository_name": "The Research Repository @ WVU (West Virginia University)",
            "source_record_id": "W3024705230",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:57",
            "updated_at": "2026-11-15 16:52:57"
        },
        {
            "id": 1004,
            "title": "Unconstrained Periocular Face Recognition: From Reconstructive Dictionary Learning to Generative Deep Learning and Beyond",
            "title_ar": "",
            "title_en": "Unconstrained Periocular Face Recognition: From Reconstructive Dictionary Learning to Generative Deep Learning and Beyond",
            "author": "Felix Juefei-Xu",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "United States",
            "governorate": "",
            "publication_year": 2018,
            "language": "en",
            "research_domain": "computer-vision",
            "research_subdomain": "face-recognition",
            "categories": "deep-learning; computer-vision; generative-ai; cnn; image-classification; face-recognition; gan",
            "keywords": "",
            "abstract": "Many real-world face recognition tasks are under unconstrained conditions such as off-angle pose variations, illumination variations, facial occlusion, facial expression, etc. In this work, we are focusing on the real-world scenarios where only the periocular region of a face is visible such as in the ISIS case. In Part I of the dissertation, we will showcase the face recognition capability based on the periocular region, which we call the periocular face recognition. We will demonstrate that face matching using the periocular region directly is more robust than the full face in terms of age-tolerant face recognition, expression-tolerant face recognition, pose-tolerant face recognition, as well as contains more cues for determining the gender information of a subject. In this dissertation, we will study direct periocular matching more comprehensively and systematically using both shallow and deep learning methods. Based on this, in Part II and Part III of the dissertation, we will continue to explore an indirect way of carrying out the periocular face recognition: periocular-based full face hallucination, because we want to capitalize on the powerful commercial face matchers and deep learning-based face recognition engines which are all trained on large-scale full face images. The reproducibility and feasibility of re-training for a proprietary facial region, such as the periocular region, is relatively low, due to the nonopen source nature of commercial face matchers as well as the amount of training data and computation power required by the deep learning based models. We will carry out the periocular-based full face hallucination based on two proposed reconstructive dictionary learning methods, including the dimensionally weighted K-SVD (DW-KSVD) dictionary learning approach and its kernel feature space counterpart using Fastfood kernel expansion approximation to reconstruct high-fidelity full face images from the periocular region, as well as two proposed generative deep learning approaches that build upon deep convolutional generative adversarial networks (DCGAN) to generate the full face from the periocular region observations, including the Gang of GANs (GoGAN) method and the discriminant nonlinear many-to-one generative adversarial networks (DNMM-GAN) for applications such as the generative open-set landmark-free frontalization (Golf) for faces and universal face optimization (UFO), which tackles an even broader set of problems than periocular based full face hallucination. Throughout Parts I-III, we will study how to handle challenging realworld scenarios such as unconstrained pose variations, unconstrained illumination conditions, and unconstrained low resolution of the periocular and facial images. Together, we aim to achieve unconstrained periocular face recognition through both direct periocular face matching and indirect periocular-based full face hallucination. In the final Part IV of the dissertation, we will go beyond and explore several new methods in deep learning that are statistically efficient for generalpurpose image recognition. Methods include the local binary convolutional neural networks (LBCNN), the perturbative neural networks (PNN), and the polynomial convolutional neural networks (PolyCNN).",
            "source_url": "https:\/\/openalex.org\/W2804539676",
            "pdf_url": "",
            "doi": "10.1184\/r1\/6724106",
            "handle": "",
            "repository_name": "KiltHub Repository",
            "source_record_id": "W2804539676",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:57",
            "updated_at": "2026-11-15 16:52:57"
        },
        {
            "id": 1042,
            "title": "Visual Aided Deep Learning Applications for Underwater Operations",
            "title_ar": "",
            "title_en": "",
            "author": "Martin Skaldebø",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "",
            "university_source": "",
            "college": "",
            "department": "",
            "country": "Norway",
            "governorate": "",
            "publication_year": 2019,
            "language": "sv",
            "research_domain": "computer-vision",
            "research_subdomain": "object-detection",
            "categories": "deep-learning; computer-vision; generative-ai; robotics; ai-applications; cnn; object-detection; gan; ai-industry",
            "keywords": "",
            "abstract": "Den 4. industrielle revolusjonen har pågått de siste fire årene og resultater har begynt å vise seg i industrimarkedet. Google slipper stadig ny teknologi som kan slå mennesker i et mangfold av strategispill og de har utviklet kunstig inteligens som kan lære å gå på egenhånd. Statlige instanser og klassifikasjonsselskaper sammarbeider mot regler og retningslinjer for autonome skip, Yara Birkeland er et skip som er planlagt ferdig i 2020 som vil gå mot full autonom operasjon innen 2022. Autonomi i undervannsoperasjoner erfarer også vekst som vises igjen i økning i markedet. Maskinlæring er forventet å være en bidragsyter i skiftet mot en mer autonom fremtid. Metoder for innhenting av relevant informasjon i undervannsapplikasjoner som eksisterer idag bruker hovedsakelig akkustiske sensorer. Metoder for informasjonsinhenting som bruker kamerabilder er sjeldent brukt, men forbedringer innen nevrale nettverk, spesielt Convolutional Neural Networks (CNN), har vist lovende resultater. Denne rapporten vil undersøke bruken av slike metoder for undervannsoperasjoner. Rapporten trekker spesielt frem to problemstillinger, som er fordelt på to vedlagte artikler. Den første problemstillingen anngår realitetsgapet, som beskriver fenomenet som oppstår når kunnskap blir overført mellom domener, mer presist mellom det kunstige simuleringsdomenet og den virkelige verden. Eksisterende metoder angående dette har blitt undersøkt og sammenlignet for å danne et utsnitt av eksisterende metoder. Grunnet det tøffe miljøet under vann har bruken av kamerabilder til informasjonsinnhenting vært liten. Dårlig belysning, maritim snø, et miljø i kontinuerlig endring etc. gjør det vanskelig å modellere et undervannsmiljø. Det vil derfor nesten garantert eksistere et realitetsgap mellom det simulerte domenet og den virkelige verden. I løpet av de siste årene har flere rammeverk som angriper realitetsgapet oppstått og blitt publisert. Generative Adversarial Networks (GAN) har muligens stått frem som en av de mest lovende metodene for domenetransformasjon de siste årene. Denne rapporten undersøker en metode av dette rammeverket kalt CycleGAN, som introduserer en sykelkonsistens egenskap i det orginale GAN rammeverket. CycleGAN har vist gode resultater inne flere disipliner og har i denne rapporten for første gang blitt anvendt på undervannsapplikasjoner. Metoden består av å ta et bilde som input og reprodusere det samme bildet i et annet domene som output, e.g. ta et generert bilde fra simuleringsdata av et undervannspanel som input og produsere et virkelighetsnært bilde av det samme undervannspanelet. I denne rapporten er rammeverket anvendt på to forskjellige dataset med bilder fra undervannsmiljøer og en god kartlegging mellom simuleringsbaserte domener og den virkelige verden er oppnådd. Den andre problemstillingen omhandler optisk basert objekt detektering ved bruk av monokularkamera. Deep learning metoder for objekt detektering er undersøkt i unison med metoder for å ekstrakte relevante egenskaper fra de respektive detektorene for å bidra i et dynamsik posisjonerings (DP) system. En detekteringsalgoritme basert på YOLOv3 er anvendt på et dataset sammensatt av bilder fra marin kybernetisk laboratoriet (MC-lab) ved NTNU. Datasettet inkluderer bilder av et kjent objekt i bassenget i laboratoriet. Når detekteringsalgoritmen er ferdig trent og kan vellykket detektere det relevante objectet, er en skaleringsfunksjon samt et DP-system for undervannsdronen BlueROV utviklet. BlueROV2 utfører DP relativ til det kjente objektet ved å innhente relevant informasjon angående posisjon og avstander ved hjelp av detekteringsalgoritmen. Systemet fungerer tilfredstillende og klarer å utføre DP med korrekt informasjonsinnhenting av både posisjon og avstander til objektet fra kamerabilder.",
            "source_url": "https:\/\/openalex.org\/W2999495304",
            "pdf_url": "",
            "doi": "",
            "handle": "",
            "repository_name": "NORA - Norwegian Open Research Archives",
            "source_record_id": "W2999495304",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:57",
            "updated_at": "2026-11-15 16:52:57"
        },
        {
            "id": 1087,
            "title": "Single-Channel Speech Enhancement Based on Deep Neural Networks",
            "title_ar": "",
            "title_en": "Single-Channel Speech Enhancement Based on Deep Neural Networks",
            "author": "Zhiheng Ouyang",
            "supervisor": "",
            "degree_type": "unspecified",
            "degree_evidence": "",
            "university": "Concordia University",
            "university_source": "repository_host",
            "college": "",
            "department": "",
            "country": "Canada",
            "governorate": "",
            "publication_year": 2019,
            "language": "en",
            "research_domain": "nlp",
            "research_subdomain": "speech-recognition",
            "categories": "deep-learning; nlp; generative-ai; cnn; speech-recognition; gan",
            "keywords": "",
            "abstract": "Speech enhancement (SE) aims to improve the speech quality of the degraded speech. Recently, researchers have resorted to deep-learning as a primary tool for speech enhancement, which often features deterministic models adopting supervised training. Typically, a neural network is trained as a mapping function to convert some features of noisy speech to certain targets that can be used to reconstruct clean speech. These methods of speech enhancement using neural networks have been focused on the estimation of spectral magnitude of clean speech considering that estimating spectral phase with neural networks is difficult due to the wrapping effect. As an alternative, complex spectrum estimation implicitly resolves the phase estimation problem and has been proven to outperform spectral magnitude estimation. In the first contribution of this thesis, a fully convolutional neural network (FCN) is proposed for complex spectrogram estimation. Stacked frequency-dilated convolution is employed to obtain an exponential growth of the receptive field in frequency domain. The proposed network also features an efficient implementation that requires much fewer parameters as compared with conventional deep neural network (DNN) and convolutional neural network (CNN) while still yielding a comparable performance. Consider that speech enhancement is only useful in noisy conditions, yet conventional SE methods often do not adapt to different noisy conditions. In the second contribution, we proposed a model that provides an automatic \"on\/off\" switch for speech enhancement. It is capable of scaling its computational complexity under different signal-to-noise ratio (SNR) levels by detecting clean or near-clean speech which requires no processing. By adopting information maximizing generative adversarial network (InfoGAN) in a deterministic, supervised manner, we incorporate the functionality of SNR-indicator into the model that adds little additional cost to the system. We evaluate the proposed SE methods with two objectives: speech intelligibility and application to automatic speech recognition (ASR). Experimental results have shown that the CNN-based model is applicable for both objectives while the InfoGAN-based model is more useful in terms of speech intelligibility. The experiments also show that SE for ASR may be more challenging than improving the speech intelligibility, where a series of factors, including training dataset and neural network models, would impact the ASR performance.",
            "source_url": "https:\/\/openalex.org\/W3199713199",
            "pdf_url": "https:\/\/spectrum.library.concordia.ca\/986846\/1\/Ouyang_MASc_F2020.pdf",
            "doi": "",
            "handle": "",
            "repository_name": "Spectrum Research Repository (Concordia University)",
            "source_record_id": "W3199713199",
            "other_sources": "",
            "verification_status": "needs_review",
            "created_at": "2026-11-15 16:52:58",
            "updated_at": "2026-11-15 16:52:58"
        }
    ]
}