My study notes of GANs specialization :)

This repo summarizes all the recommended reading materials and notebooks that are related to GANs offered by DeeplearningAI. The reading materials are not all covered in details in the specialization courses, but some notebooks and recommended readings could be used as a starting point to learn different algorithms and applications.

• Hyperspherical Variational Auto-Encoders (Davidson, Falorsi, De Cao, Kipf, and Tomczak, 2018): https://www.researchgate.net/figure/Latent-space-visualization-of-the-10-MNIST-digits-in-2-dimensions-of-both-N-VAE-left_fig2_324182043
• Analyzing and Improving the Image Quality of StyleGAN (Karras et al., 2020): https://arxiv.org/abs/1912.04958
• Semantic Image Synthesis with Spatially-Adaptive Normalization (Park, Liu, Wang, and Zhu, 2019): https://arxiv.org/abs/1903.07291
• Few-shot Adversarial Learning of Realistic Neural Talking Head Models (Zakharov, Shysheya, Burkov, and Lempitsky, 2019): https://arxiv.org/abs/1905.08233
• Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling (Wu, Zhang, Xue, Freeman, and Tenenbaum, 2017): https://arxiv.org/abs/1610.07584
• These Cats Do Not Exist (Glover and Mott, 2019): http://thesecatsdonotexist.com/
• Large Scale GAN Training for High Fidelity Natural Image Synthesis (Brock, Donahue, and Simonyan, 2019): https://arxiv.org/abs/1809.11096
• PyTorch Documentation: https://pytorch.org/docs/stable/index.html#pytorch-documentation
• MNIST Database: http://yann.lecun.com/exdb/mnist/

• Transposed convolutions : Odena, et al., "Deconvolution and Checkerboard Artifacts", Distill, 2016. http://doi.org/10.23915/distill.00003
• DCGAN paper : Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks (Radford, Metz, and Chintala, 2016): https://arxiv.org/abs/1511.06434
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks (Radford, Metz, and Chintala, 2016): https://arxiv.org/abs/1511.06434

• In this notebook, you're going to learn about TGAN, from the paper Temporal Generative Adversarial Nets with Singular Value Clipping (Saito, Matsumoto, & Saito, 2017), and its origins in image generation. Notebook link: https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C1W2_Video_Generation_(Optional).ipynb

• Spectral normalization (SNGAN), a weight normalization technique to stabilize the training of the discriminator, as proposed in Spectral Normalization for Generative Adversarial Networks (Miyato et al. 2018).
• Protein GAN: https://www.biorxiv.org/content/10.1101/789719v2
• WGAN Wasserstein GAN (Arjovsky, Chintala, and Bottou, 2017): https://arxiv.org/abs/1701.07875
• WGAN-GP Improved Training of Wasserstein GANs (Gulrajani et al., 2017): https://arxiv.org/abs/1704.00028 Gradient penalty (GP) as well as weight clipping to WGAN in order to enforce 1-Lipschitz continuity and improve stability
• From GAN to WGAN (Weng, 2017): https://lilianweng.github.io/lil-log/2017/08/20/from-GAN-to-WGAN.html This article provides a great walkthrough of how WGAN addresses the difficulties of training a traditional GAN with a focus on the loss functions.

• ProteinGAN: https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/ProteinGAN.ipynb

• Conditional GAN: Conditional Generative Adversarial Nets (Mirza and Osindero, 2014): https://arxiv.org/abs/1411.1784
• An example of Controllable GAN: Interpreting the Latent Space of GANs for Semantic Face Editing (Shen, Gu, Tang, and Zhou, 2020): https://arxiv.org/abs/1907.10786

• A Note on the Inception Score (Barratt and Sharma, 2018): https://arxiv.org/abs/1801.01973 Know more about why Fréchet Inception Distance has overtaken Inception Score, this paper illustrates the problems with using Inception Score.
• HYPE: A Benchmark for Human eYe Perceptual Evaluation of Generative Models (Zhou et al., 2019): https://arxiv.org/abs/1904.01121 Human evaluation and HYPE (Human eYe Perceptual Evaluation) of GANs.
• Improved Precision and Recall Metric for Assessing Generative Models (Kynkäänniemi, Karras, Laine, Lehtinen, and Aila, 2019): https://arxiv.org/abs/1904.06991
• Fréchet Inception Distance (Jean, 2018): https://nealjean.com/ml/frechet-inception-distance/
• GAN — How to measure GAN performance? (Hui, 2018): https://medium.com/@jonathan_hui/gan-how-to-measure-gan-performance-64b988c47732

• Machine Bias (Angwin, Larson, Mattu, and Kirchner, 2016): https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
• Fairness Definitions Explained (Verma and Rubin, 2018): https://fairware.cs.umass.edu/papers/Verma.pdf
• Machine Learning Glossary: Fairness (2020): https://developers.google.com/machine-learning/glossary/fairness
• A Survey on Bias and Fairness in Machine Learning (Mehrabi, Morstatter, Saxena, Lerman, and Galstyan, 2019): https://arxiv.org/abs/1908.09635
• Does Object Recognition Work for Everyone? (DeVries, Misra, Wang, and van der Maaten, 2019): https://arxiv.org/abs/1906.02659
• What a machine learning tool that turns Obama white can (and can't) tell us about AI bias (Vincent, 2020): https://www.theverge.com/21298762/face-depixelizer-ai-machine-learning-tool-pulse-stylegan-obama-bias
• Fair Attribute Classification through Latent Space De-biasing. Vikram V. Ramaswamy, Sunnie S. Y. Kim, Olga Russakovsky. CVPR 2021.
• Hyperspherical Variational Auto-Encoders (Davidson, Falorsi, De Cao, Kipf, and Tomczak, 2018): https://arxiv.org/abs/1804.00891
• Generating Diverse High-Fidelity Images with VQ-VAE-2 (Razavi, van den Oord, and Vinyals, 2019): https://arxiv.org/abs/1906.00446
• Conditional Image Generation with PixelCNN Decoders (van den Oord et al., 2016): https://arxiv.org/abs/1606.05328
• Glow: Better Reversible Generative Models (Dhariwal and Kingma, 2018): https://openai.com/blog/glow/
• PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models (Menon, Damian, Hu, Ravi, and Rudin, 2020): https://arxiv.org/abs/2003.03808

• VAE
• Score-based GAN model: https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C2W2_(Optional_Notebook)_Score_Based_Generative_Modeling.ipynb This is a hitchhiker's guide to score-based generative models, a family of approaches based on estimating gradients of the data distribution. They have obtained high-quality samples comparable to GANs (like below, figure from this paper) without requiring adversarial training, and are considered by some to be the new contender to GANs.
• GAN debiasing: https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C2W2_GAN_Debiasing_(Optional).ipynb In this notebook, you will learn about Fair Attribute Classification through Latent Space De-biasing (Ramaswamy et al. 2020) that introduces a method for training accurate target classifiers while mitigating biases that stem from these correlations. Specifically, this work uses GANs to generate realistic-looking images and perturb these images in the underlying latent space to generate training data that is balanced for each protected attribute. They augment the original dataset with this perturbed generated data, and empirically demonstrate that target classifiers trained on the augmented dataset exhibit a number of both quantitative and qualitative benefits.
• NeRF: Neural Radiance Fields: https://colab.research.google.com/drive/18DladhUz7_U8iBkkQxMBk2f7C2NAvPCC?usp=sharing In this notebook, you'll learn how to use Neural Radiance Fields to generate new views of a complex 3D scene using only a couple input views, first proposed by NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis (Mildenhall et al. 2020). Though 2D GANs have seen success in high-resolution image synthesis, NeRF has quickly become a popular technique to enable high-resolution 3D-aware GANs.

• StyleGAN: A Style-Based Generator Architecture for Generative Adversarial Networks (Karras, Laine, and Aila, 2019): https://arxiv.org/abs/1812.04948
• GAN — StyleGAN & StyleGAN2 (Hui, 2020): https://medium.com/@jonathan_hui/gan-stylegan-stylegan2-479bdf256299
• Generative Adversarial Networks (Goodfellow et al., 2014): https://arxiv.org/abs/1406.2661
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks (Radford, Metz, and Chintala, 2016): https://arxiv.org/abs/1511.06434
• Coupled Generative Adversarial Networks (Liu and Tuzel, 2016): https://arxiv.org/abs/1606.07536
• Progressive Growing of GANs for Improved Quality, Stability, and Variation (Karras, Aila, Laine, and Lehtinen, 2018): https://arxiv.org/abs/1710.10196
• The Unusual Effectiveness of Averaging in GAN Training (Yazici et al., 2019): https://arxiv.org/abs/1806.04498v2
• StyleGAN - Official TensorFlow Implementation (Karras et al., 2019): https://github.com/NVlabs/stylegan
• StyleGAN Faces Training (Branwen, 2019): https://www.gwern.net/images/gan/2019-03-16-stylegan-facestraining.mp4
• Facebook AI Proposes Group Normalization Alternative to Batch Normalization (Peng, 2018): https://medium.com/syncedreview/facebook-ai-proposes-group-normalization-alternative-to-batch-normalization-fb0699bffae7

• In this notebook, you're going to learn about StyleGAN2, from the paper Analyzing and Improving the Image Quality of StyleGAN (Karras et al., 2019), and how it builds on StyleGAN. This is the V2 of StyleGAN, so be prepared for even more extraordinary outputs. a relative link
• In this notebook, you'll learn about and implement the components of BigGAN, the first large-scale GAN architecture proposed in Large Scale GAN Training for High Fidelity Natural Image Synthesis (Brock et al. 2019). BigGAN performs a conditional generation task, so unlike StyleGAN, it conditions on a certain class to generate results. BigGAN is based mainly on empirical results and shows extremely good results when trained on ImageNet and its 1000 classes. a relative link

• Use GANs to create talking heads and deepfakes: Few-Shot Adversarial Learning of Realistic Neural Talking Head Models (Zakharov, Shysheya, Burkov, and Lempitsky, 2019): https://arxiv.org/abs/1905.08233
• De-identify (anonymize) a face while preserving essential facial attributes in order to conceal an identity: De-identification without losing faces (Li and Lyu, 2019): https://arxiv.org/abs/1902.04202
• Attributing Fake Images to GANs: Learning and Analyzing GAN Fingerprints (Yu, Davis, and Fritz, 2019): https://arxiv.org/abs/1811.08180

• Generative Teaching Network (GTN), first introduced in Generative Teaching Networks: Accelerating Neural Architecture Search by Learning to Generate Synthetic Training Data (Such et al. 2019). Essentially, a GTN is composed of a generator (i.e. teacher), which produces synthetic data, and a student, which is trained on this data for some task. The key difference between GTNs and GANs is that GTN models work cooperatively (as opposed to adversarially). https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C3W1_Generative_Teaching_Networks_(Optional).ipynb
• Data Augmentation, a relative link.

• Image-to-Image Translation with Conditional Adversarial Networks (Isola, Zhu, Zhou, and Efros, 2018): https://arxiv.org/abs/1611.07004
• Patch-Based Image Inpainting with Generative Adversarial Networks (Demir and Unal, 2018): https://arxiv.org/abs/1803.07422

• In this notebook, you will learn about Pix2PixHD, which synthesizes high-resolution images from semantic label maps. Proposed in High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs (Wang et al. 2018), Pix2PixHD improves upon Pix2Pix via multiscale architecture, improved adversarial loss, and instance maps. https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C3W2_Pix2PixHD_(Optional).ipynb
• In this notebook, you will learn about Super-Resolution GAN (SRGAN), a GAN that enhances the resolution of images by 4x, proposed in Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network (Ledig et al. 2017). You will also implement the architecture and training in full and be able to train it on the CIFAR dataset. https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C3W2_SRGAN_(Optional).ipynb
• In this notebook, you will learn about GauGAN, which synthesizes high-resolution images from semantic label maps, which you implement and train. GauGAN is based around a special denormalization technique proposed in Semantic Image Synthesis with Spatially-Adaptive Normalization (Park et al. 2019). https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C3W2_GauGAN_(Optional).ipynb
• UNet, a relative link.
• Pix2Pix, a relative link.

• CycleGAN, Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks (Zhu, Park, Isola, and Efros, 2020): https://arxiv.org/abs/1703.10593
• Data augmentation using generative adversarial networks (CycleGAN) to improve generalizability in CT segmentation tasks (Sandfort, Yan, Pickhardt, and Summers, 2019): https://www.nature.com/articles/s41598-019-52737-x.pdf
• Distribution Matching Losses Can Hallucinate Features in Medical Image Translation (Cohen, Luck, and Honari, 2018): https://arxiv.org/abs/1805.08841

• In this notebook, you will learn about and implement MUNIT, a method for unsupervised image-to-image translation, as proposed in Multimodal Unsupervised Image-to-Image Translation (Huang et al. 2018). https://colab.research.google.com/github/https-deeplearning-ai/GANs-Public/blob/master/C3W3_MUNIT_(Optional).ipynb
• PyTorch implementation of CycleGAN (2017): https://github.com/togheppi/CycleGAN
• Unsupervised Image-to-Image Translation (NVIDIA, 2018): https://github.com/mingyuliutw/UNIT
• Multimodal Unsupervised Image-to-Image Translation (Huang et al., 2018): https://github.com/NVlabs/MUNIT
• CycleGAN a relative link.