{ "exported_at": "2026-11-15 18:24:21", "count": 8, "records": [ { "id": 143, "title": "Artificial Intelligence and Cybersecurity: Building an Automotive Cybersecurity Framework Using Machine Learning Algorithms", "title_ar": "", "title_en": "Artificial Intelligence and Cybersecurity: Building an Automotive Cybersecurity Framework Using Machine Learning Algorithms", "author": "Nevrus Kaja", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2019, "language": "en", "research_domain": "responsible-ai", "research_subdomain": "fairness", "categories": "machine-learning; robotics; responsible-ai; expert-systems; ai-applications; fairness; ai-cybersecurity", "keywords": "", "abstract": "Automotive technology has continued to advance in many aspects. As an outcome of such advancements, autonomous vehicles are closer to commercialization and have brought to life a complex automotive technology ecosystem [1]. Like every other technology, these developments bring benefits but also introduce a variety of risks. One of these risks in the automotive space is cybersecurity threats. In the case of cars, these security challenges can produce devastating results and tremendous costs, including loss of life. Therefore, conducting a clear analysis, assessment and detection of threats solves some of the cybersecurity challenges in the automotive ecosystem. This dissertation does just that, by building a three-step framework to analyze, assess,and detect threats using machine learning algorithms. First, it does an analysis of the connected vehicle threats while leveraging the STRIDE framework [2]. Second, it presents an innovative, Fuzzy based threat assessment model (FTAM). FTAM leverages threat characterizations from established threat assessment models while focusing on improving its assessment capabilities by using Fuzzy logic. Through this methodology, FTAM can improve the efficiency and accuracy of the threat assessment process by using Fuzzy logic to determine the “degree” of the threat over other existing methods. This differs from the current threat assessment models which use subjective assessment processes based on table look-ups or scoring. Thirdly, this dissertation proposes an intrusion detection system (IDS) to detect malicious threats while taking in consideration results from the previous assessment stage. This IDS uses the dataset provided from Wyoming Connected Vehicle Deployment program [3] and consists of a two-stage intrusion detection system based on supervised and unsupervised machine learning algorithms. The first stage uses unsupervised learning to detect whether there is an attack present and the second stage classifies these attacks in a supervised learning fashion. The second stage also addresses data bias and eliminates the number of false positives. The simulation of this approach results in an IDS able to detect and classify attacks at a 99.965% accuracy and lowers the false positives rate to 0%.", "source_url": "https:\/\/openalex.org\/W2950711294", "pdf_url": "https:\/\/deepblue.lib.umich.edu\/bitstreams\/379b5e3d-3843-4c2a-abfa-84d77c90685e\/download", "doi": "", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W2950711294", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:47", "updated_at": "2026-11-15 16:52:47" }, { "id": 344, "title": "Deep Learning Approaches for Condition Monitoring and Prognostic Analysis", "title_ar": "", "title_en": "Deep Learning Approaches for Condition Monitoring and Prognostic Analysis", "author": "Zunya Shi", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2021, "language": "en", "research_domain": "medical-ai", "research_subdomain": "predictive-analytics", "categories": "machine-learning; deep-learning; medical-ai; ai-applications; predictive-analytics; ai-industry", "keywords": "", "abstract": "Predictive analytics is a technique to make predictions about future unknown events by analyzing current and historical data. The market for predictive analysis is expected to reach $12.4 billion in 2022. Prognostic analysis is one dimension of predictive analytics in engineering discipline and focuses on predicting the time when a system or a component will no longer perform its intended function. A robust and accurate prognostic analysis makes for the increment of production reliability and safety as well as maintenance cost reduction, thus becoming one key task of condition-based maintenance (CBM). Given multiple sensors to simultaneously monitor the health condition during the age of Internet of Things (IoT) and Industry 4.0, traditional prognostic technologies have the limited ability of handling large-scale dataset embedded with complex structures. Therefore, this dissertation mainly focuses on developing high-performance Deep Learning (DL) technologies for condition monitoring and prognostic analysis under the data-rich environment. Besides the CBM, the proposed DL technologies can be also successfully extended and applied into other fields, like battery power estimation, Alzheimer’s Disease (AD) prediction, etc., which have been validated in the dissertation with real case studies. In addition, this dissertation also studied and answered research questions like how to evaluate the quality of features for prognostic analysis, how to select and extract high-level features from original dataset, how to capture the interactions between different features, and so forth. The answers for these questions play important roles in model development for prognostic analysis.", "source_url": "https:\/\/openalex.org\/W4287024228", "pdf_url": "", "doi": "10.7302\/2533", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W4287024228", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:48", "updated_at": "2026-11-15 16:52:48" }, { "id": 349, "title": "Deep Learning and Reward Design for Reinforcement Learning", "title_ar": "", "title_en": "Deep Learning and Reward Design for Reinforcement Learning", "author": "Xiaoxiao Guo", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2017, "language": "en", "research_domain": "nlp", "research_subdomain": "cnn", "categories": "deep-learning; computer-vision; nlp; reinforcement-learning; cnn", "keywords": "", "abstract": "One of the fundamental problems in Artificial Intelligence is sequential decision making in a flexible environment. Reinforcement Learning (RL) gives a set of tools for solving sequential decision problems. Although the theory of RL addresses a general class of learning problems with a constructive mathematical formulation, the challenges posed by the interaction of rich perception and delayed rewards in many domains remain a significant barrier to the widespread applicability of RL methods. The rich perception problem itself has two components: 1) the sensors at any time step do not capture all the information in the history of observations, leading to partial observability, and 2) the sensors provide very high-dimensional observations, such as images and natural languages, that introduce computational and sample-complexity challenges for the representation and generalization problems in policy selection. The delayed reward problem—that the effect of actions in terms of future rewards is delayed in time—makes it hard to determine how to credit action sequences for reward outcomes. This dissertation offers a set of contributions that adapt the hierarchical representation learning power of deep learning to address rich perception in vision and text domains, and develop new reward design algorithms to address delayed rewards. The first contribution is a new learning method for deep neural networks in vision-based real-time control. The learning method distills slow policies of the Monte Carlo Tree Search (MCTS) into fast convolutional neural networks, which outperforms the conventional Deep Q-Network. The second contribution is a new end-to-end reward design algorithm to mitigate the delayed rewards for the state-of-the-art MCTS method. The reward design algorithm converts visual perceptions into reward bonuses via deep neural networks, and optimizes the network weights to improve the performance of MCTS end-to-end via policy gradient. The third contribution is to extend existing policy gradient reward design method from single task to multiple tasks. Reward bonuses learned from old tasks are transferred to new tasks to facilitate learning. The final contribution is an application of deep reinforcement learning to another type of rich perception, ambiguous texts. A synthetic data set is proposed to evaluate the querying, reasoning and question-answering abilities of RL agents, and a deep memory network architecture is applied to solve these challenging problems to substantial degrees.", "source_url": "https:\/\/openalex.org\/W2751395930", "pdf_url": "", "doi": "", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W2751395930", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:48", "updated_at": "2026-11-15 16:52:48" }, { "id": 463, "title": "Language Supervision for Computer Vision", "title_ar": "", "title_en": "Language Supervision for Computer Vision", "author": "Karan Desai", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2024, "language": "en", "research_domain": "nlp", "research_subdomain": "llm", "categories": "machine-learning; deep-learning; computer-vision; nlp; transfer-learning; few-zero-shot; multimodal; object-detection; image-segmentation; llm", "keywords": "", "abstract": "Representation learning lies at the core of modern Artificial Intelligence. In computer vision, labeled image datasets like ImageNet have been the standard choice for representation learning. Despite being empirically successful, this approach is expensive to scale due to labeling costs. Moreover, the representation quality is limited by the size and diversity of datasets and their associated label ontologies. My research explores using natural language supervision for computer vision. Using natural language allows us to go beyond fixed label ontologies and scale up to more general sources such as internet data. Toward this goal, my dissertation explores four problems -- (1) Learning representations: I propose one of the first methods for language-supervised visual learning that uses image captioning as the training objective, showing its efficacy compared to ImageNet-trained methods on downstream tasks like object detection and segmentation. (2) Scaling data: I explore social media as a rich source of high-quality image descriptions and curate a dataset of 12 million image-text pairs while ensuring responsible curation practices. (3) Understanding data: It is difficult to comprehend the diversity of visual concepts present in millions of image-text pairs. I posit that images and text naturally organize into a tree-like hierarchy and propose an approach for learning representations that capture this hierarchy using tools from hyperbolic geometry. (4) Transfer to downstream tasks: Large vision-language models show impressive zero-shot transfer capabilities on image-level tasks like classification and retrieval. However, their transferability to pixel-level tasks like object detection and segmentation has relied on expensive labeled mask annotations. I propose an object detector to efficiently transfer pre-trained vision models to segment and classify visual objects without any fine-tuning, unlike existing detectors that train using orders of magnitude more labeled masks to achieve high performance. In summary, my research affirms that using language supervision can drive the next leap of progress in computer vision and has immense utility in practical applications.", "source_url": "https:\/\/openalex.org\/W4398230580", "pdf_url": "", "doi": "10.7302\/22865", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W4398230580", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:49", "updated_at": "2026-11-15 16:52:49" }, { "id": 555, "title": "Demographic-Aware Natural Language Processing", "title_ar": "", "title_en": "Demographic-Aware Natural Language Processing", "author": "Aparna Garimella", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2020, "language": "en", "research_domain": "nlp", "research_subdomain": "speech-recognition", "categories": "machine-learning; nlp; predictive-analytics; speech-recognition", "keywords": "", "abstract": "The underlying traits of our demographic group affect and shape our thoughts, and therefore surface in the way we express ourselves and employ language in our day-to-day life. Understanding and analyzing language use in people from different demographic backgrounds help uncover their demographic particularities. Conversely, leveraging these differences could lead to the development of better language representations, thus enabling further demographic-focused refinements in natural language processing (NLP) tasks. In this thesis, I employ methods rooted in computational linguistics to better understand various demographic groups through their language use. The thesis makes two main contributions. First, it provides empirical evidence that words are indeed used differently by different demographic groups in naturally occurring text. Through experiments conducted on large datasets which display usage scenarios for hundreds of frequent words, I show that automatic classification methods can be effective in distinguishing between word usages of different demographic groups. I compare the encoding ability of the utilized features by conducting feature analyses, and shed light on how various attributes contribute to highlighting the differences. Second, the thesis explores whether demographic differences in word usage by different groups can inform the development of more refined approaches to NLP tasks. Specifically, I start by investigating the task of word association prediction. The thesis shows that going beyond the traditional ``one-size-fits-all'' approach, demographic-aware models achieve better performances in predicting word associations for different demographic groups than generic ones. Next, I investigate the impact of demographic information on part-of-speech tagging and syntactic parsing, and the experiments reveal numerous part-of-speech tags and syntactic relations, whose predictions benefit from the prevalence of a specific group in the training data. Finally, I explore demographic-specific humor generation, and develop a humor generation framework to fill-in the blanks to generate funny stories, while taking into account people's demographic backgrounds.", "source_url": "https:\/\/openalex.org\/W3139872221", "pdf_url": "https:\/\/deepblue.lib.umich.edu\/bitstreams\/de2aac6e-ca29-4c11-b396-29d819be092d\/download", "doi": "", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W3139872221", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:50", "updated_at": "2026-11-15 16:52:50" }, { "id": 631, "title": "Discovery in Reinforcement Learning", "title_ar": "", "title_en": "Discovery in Reinforcement Learning", "author": "Vivek Veeriah", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2022, "language": "en", "research_domain": "computer-vision", "research_subdomain": "predictive-analytics", "categories": "machine-learning; deep-learning; computer-vision; reinforcement-learning; robotics; predictive-analytics", "keywords": "", "abstract": "This thesis focuses on Reinforcement Learning (RL) which considers an agent that makes sequen- tial decisions in an environment to maximize its reward. Recent breakthroughs on domains such as Go, Atari games, and Robotics are a result of combining advances from Deep Learning with RL algorithms. This combination of RL with deep neural networks has moved the field from hand- designing feature representations to learning features directly from raw data (e.g., pixel images). Despite these advancements, RL practitioners still need to define what predictive knowledge an agent needs to learn, instead of the agent autonomously discovering such knowledge. Developing the ability to discover what predictive knowledge is useful to learn about could allow the RL agents to learn more efficiently. In this thesis, we identify some of the challenges towards designing agents with the aforementioned ability, and propose and evaluate methods to address them. In the first part of this thesis, we present an approach for discovering multiple subgoals to learn optimal policies from a single stream of experience. We show that this approach is a useful pre- training procedure and as a source of auxiliary task updates when there is a main task of interest. In the second part of this thesis, we present three approaches to discover useful knowledge of various forms to maximize the agent’s task performance. The first one introduces an architecture and an associated meta-gradient algorithm to discover predictive questions about the agent’s ex- perience to drive representation learning in RL agents. We show that it helps in faster learning when used as auxiliary tasks. The second work introduces another meta-gradient approach to dis- cover temporal-abstractions in the form of options that enables a hierarchical RL agent to learn faster on new, unseen tasks. The third one presents an algorithm to select a small number of affor- dances in the form of actions or options from a continuous space to improve the performance of a model-based planning agent on hierarchical tasks. The final part studies the problem of learning and maintaining accurate knowledge in the form of option-conditional predictions in a partially-observable environment. Specifically, we implement and empirically demonstrate the thought experiment by Ring [2021], which provided a detailed blueprint of how high-level, abstract knowledge could potentially be represented as layered pre- dictions of agent’s sensorimotor stream in navigation environments. In addition to this empirical demonstration, we introduce an approach to discover those option-conditional predictions which were previously hand-defined and demonstrate the feasibility of this approach.", "source_url": "https:\/\/openalex.org\/W4294885165", "pdf_url": "", "doi": "10.7302\/6337", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W4294885165", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:51", "updated_at": "2026-11-15 16:52:51" }, { "id": 657, "title": "Design and Implementation of an Autonomous Driving System with a Deep Learning Approach on a Scaled Vehicle Platform", "title_ar": "", "title_en": "Design and Implementation of an Autonomous Driving System with a Deep Learning Approach on a Scaled Vehicle Platform", "author": "Jesudara Omidokum", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2022, "language": "en", "research_domain": "deep-learning", "research_subdomain": "ai-education", "categories": "deep-learning; robotics; ai-applications; ai-education", "keywords": "", "abstract": "The thesis project is developed to create a platform for autonomous driving research and education purposes. The main goal of this study is to present a complete and adequate way for building a platform with a drive-by-wire system and sensor packages that are both comprehensive and appropriate for testing and deploying machine-learning-based algorithms on a scaled vehicle. The vehicle platform addresses reproducibility issues, onboard processing capability restrictions, vehicle scale, and system cost with the existing related vehicle platforms. Our vehicle platform system integrates both drives-by-wire and an autonomous system enabled with sensor packages in an easy-to-implement format. The platform is a generic and multilevel system, with flight controller programs combined with a GPS module handling the mid to high-level motor controls and a laptop powered by a graphics processing unit (GPU) capable of handling the advanced and more complex algorithms. The vehicle platform is validated by employing it in a deep-learning-based behavioral cloning study. The platform's affordability and adaptability would benefit broader research and the education community.", "source_url": "https:\/\/openalex.org\/W4311648930", "pdf_url": "", "doi": "10.7302\/6650", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W4311648930", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:52:51", "updated_at": "2026-11-15 16:52:51" }, { "id": 1236, "title": "Explorations of In-Context Reinforcement Learning", "title_ar": "", "title_en": "Explorations of In-Context Reinforcement Learning", "author": "Ethan Brooks", "supervisor": "", "degree_type": "unspecified", "degree_evidence": "", "university": "University of Michigan", "university_source": "repository_host", "college": "", "department": "", "country": "United States", "governorate": "", "publication_year": 2024, "language": "en", "research_domain": "nlp", "research_subdomain": "llm", "categories": "machine-learning; deep-learning; nlp; reinforcement-learning; self-semi-supervised; rnn; transformers; llm", "keywords": "", "abstract": "In-Context Learning describes a form of learning that occurs when a model accumulates information in its context or memory. For example, a Long Short-Term Memory unit (LSTM) can rapidly adapt to a novel task as input\/target exemplars are fed into it. While in-context learning of this kind is not a new discovery, recent work has demonstrated the capacity of large \"foundation\" models to acquire this ability \"for free,\" by training on large quantities of semi-supervised data, without the sophisticated (but often unstable) meta-objectives proposed by many earlier papers #cite(< finn2017model>)#cite(< stadie2018some>)#cite(< rakelly2019efficient>). In this work we explore several algorithms which specialize in-context learning based on semi-supervised methods to the reinforcement learning (RL) setting. In particular, we explore three approaches to in-context learning of value (expected cumulative discounted reward). The first of these methods demonstrates a method for implementing policy iteration, a classic RL algorithm, using a pre-trained large language model (LLM). We use the LLM to generate planning rollouts and extract monte-carlo estimates of value from them. We demonstrate the method on several small, text-based domains and present evidence that the LLM can generalize to unseen states, a key requirement of learning in non-tabular settings. The second method imports many of the ideas of the first, but trains a transformer model directly on offline RL data. We incorporate Algorithm Distillation (AD) #cite(< laskin2022context>), another method for in-context reinforcement learning that directly distills the improvement operator from data that includes behavior ranging from random to optimal. Our method combines the benefits of AD with the policy iteration method proposed in our previous work and demonstrates benefits in performance and generalization. Our third method proposes a new method for estimating value. Like the previous methods, this one implements a form of policy iteration, but eschews monte-carlo rollouts for a new approach to estimating value. We train a network to estimate Bellman updates and iteratively feed its outputs back into itself until the estimate converges. We find that this iterative approach improves the capability of the value estimates to generalize and mitigates some of the instability of other offline methods.", "source_url": "https:\/\/openalex.org\/W4398230562", "pdf_url": "", "doi": "10.7302\/23097", "handle": "", "repository_name": "Deep Blue (University of Michigan)", "source_record_id": "W4398230562", "other_sources": "", "verification_status": "needs_review", "created_at": "2026-11-15 16:53:00", "updated_at": "2026-11-15 16:53:00" } ] }