CVPR18: Tutorial: Part 4: Generative Adversarial Networks

Organizers: Jun-Yan Zhu
Taesung Park
Mihaela Rosca
Phillip Isola
Ian Goodfellow.
Description: Generative adversarial networks (GANs) have been at the forefront of research on generative models in the last couple of years. GANs have been used for image genera-tion, image processing, image synthesis from captions, image editing, visual domain adaptation, data generation for visual recognition, and many other applications, often leading to state of the art results. This tutorial aims to provide a broad overview of generative adversarial networks, mainly including the following three parts: (1) theoretical foundations such as basic concepts, mechanisms, and theoretical considerations, (2) best practices of the current state-of-the-art GAN and conditional GAN models, including network architectures, objective functions, and other training tricks, and (3) computer vision applications including visual domain adaptation, image processing (e.g., restoration, inpainting, super-resolution), image synthesis and manipulation, video prediction and gener-ation, 3D modeling, synthetic data generation for visual recog-nition, robotic learning, and so on.
Introduction to Generative Adversarial Networks, Ian Goodfellow (Google Brain)
Paired Image-to-Image Translation, Phillip Isola (MIT)
Unpaired Image-to-Image Translation, Taesung Park (UC Berkeley and Jun-Yan Zhu, MIT)

Can GANs Actually Learn the Distribution? Some Obstacles, Sanjeev Arora (Princeton)
TBA, Emily Denton (NYU)
Autoencoder, VAE, and GANs, Mihaela Rosca (Deepmind)
TBA, Ming-Yu Liu (NVIDIA)
Adversarial Domain Adaptation, Judy Hoffman (UC Berkeley)
TBA, Abhinav Gupta (CMU)
Generative Adversarial Imitation Learning, Stefano Ermon (Stanford)
1Video Generation and Prediction, Carl Vondrick (Columbia Univ.)
TBA, Alexei A. Efros (UC Berkeley)