Well-known references for ML and GeoSciML

Classical DL Architectures

Computer Vision

Historical paper

• Learning representations by back-propagating errors, 1986

• Gradient-based learning applied to document recognition

Convolutional neural network (CNNs):

• ImageNet Classification with Deep Convolutional Neural Networks — AlexNet (Krizhevsky et al., 2012)

• Very Deep Convolutional Networks for Large-Scale Image Recognition — VGG (Simonyan & Zisserman, 2014)

• Deep Residual Learning for Image Recognition — ResNet (He et al., 2015)

• Rethinking the Inception Architecture for Computer Vision — Inception (Szegedy et al., 2015)

Derived CNNs:

• U-Net: Convolutional Networks for Biomedical Image Segmentation — Segmentation

• You Only Look Once — YOLO (Object Detection)

• Mask R-CNN — Instance Segmentation

• ViT - Vision Transformers

• [ConvNeXt](A ConvNet for the 2020s)

• EfficientNet

Generative methods

Autoencoders & VAE:

• Auto-Encoding Variational Bayes

• An Introduction to Variational Autoencoders

Generative Adversarial Networks:

• Generative Adversarial Networks — GAN (Goodfellow et al., 2014)

• Conditional Generative Adversarial Nets — cGAN

Diffusion Models:

• Diffusion Models: A Comprehensive Survey of Methods and Applications

• Denoising Diffusion Probabilistic Models — DDPM (Ho et al., 2020)

• High-Resolution Image Synthesis with Latent Diffusion Models — Stable Diffusion

Flow Matching

• Flow Matching for Generative Modeling

• Flow Matching Guide and Code

• SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers

Datasets

GeoSciML data challenges and benchmarks