Encyclopedia of Modern Deep Learning Architectures

From multilayer perceptrons to transformers, state space models, and graph, spiking and physics-informed networks: each architecture has an interactive diagram, equations explained symbol by symbol, tested PyTorch code, and documented references with links.

  • 72architectures
  • 9categories
  • 162explained equations
  • 7interactive labs
  • 7learning paths
  • 143documented references

Architecture categories

Featured architectures

Advanced2024

Kolmogorov–Arnold Networks

Kolmogorov–Arnold Networks (KAN)

Networks that place learnable activation functions (B-spline curves) on the edges instead of fixed weights, while the nodes only sum their inputs, inspired by the Kolmog…

Advanced Neural Architectures

Intermediate2017

Capsule Networks

Capsule Networks (CapsNet)

Networks that represent entities as capsules (vectors) whose length expresses the probability that the entity is present and whose direction expresses its properties (po…

Advanced Neural Architectures

Intermediate1997

Spiking Neural Networks

Spiking Neural Networks (SNN)

Networks whose neurons communicate with discrete binary spikes in time, like the brain. The membrane potential accumulates and fires a spike when it exceeds the threshol…

Advanced Neural Architectures

Basic2016

Residual Network

Residual Network (ResNet)

It adds skip connections that make each block learn the residual difference F(x) instead of the full transformation, so training networks with hundreds of layers becomes…

Foundations

Basic2015

U-Net

U-Net

A U-shaped segmentation architecture: a contracting path that captures context and an expanding path that restores spatial resolution, with skip connections that carry f…

FoundationsHybrid Architectures

Intermediate2014

Generative Adversarial Networks

Generative Adversarial Networks (GAN)

A generative framework of two competing networks: a generator that turns random noise into samples, and a discriminator that tries to tell real samples from generated on…

Generative Networks

Intermediate2014

Variational Autoencoder

Variational Autoencoder (VAE)

A probabilistic generative model with a latent variable: an encoder that estimates an approximate distribution q(z x) and a decoder that represents p(x z), trained toget…

Generative Networks

Advanced2020

Diffusion Models

Diffusion Models (DDPM)

They gradually corrupt the data by adding Gaussian noise over T steps, then a network learns to reverse this process by predicting the added noise, so a new example is g…

Generative Networks

Advanced2022

Latent Diffusion Models

Latent Diffusion Models (LDM)

They apply diffusion in a compressed latent space from a pretrained autoencoder instead of pixel space, with a cross-attention mechanism for feeding in conditions (text,…

Generative NetworksHybrid Architectures

All architectures

Interactive labs

Learning paths

How was the encyclopedia prepared?

  • Every reference is linked to its official identifier (DOI, arXiv or the publisher's page), and its metadata was verified automatically; the few references without an identifier are marked as such.
  • Every PyTorch code sample in the encyclopedia was actually run on small synthetic data, with the run date, duration and outputs shown below it. These are simplified teaching examples, not the papers' original implementations.
  • The diagrams are illustrations with the dimensions of a small example, not drawings matching every detail of the original papers.
  • Each page has a review status and date: “Draft”, “Scientific review” or “Published”. Current pages are in “Scientific review” status until a specialist reviews them.
  • We do not describe an architecture as “the best” or “the newest” without a criterion and a date: performance depends on the data, the task and the resources.