Advanced Topics In Computer Vision And Deep Learning
Spring 2019



Course Schedule

Date Topic Lecturer Slides Papers
24/3 Introduction and Deep Internal Learning
Michal & Assaf Intro Zero Shot Super-Resolution using Deep Internal Learning (CVPR'18)
Double-DIP: Unsupervised Image Decomposition via Coupled Deep-Image-Priors (CVPR'19)
InGAN: Capturing and Remapping the "DNA" of a Natural Image
31/3 Advances in GANs
Sefi Bell Kligler & Akhiad Bercovich GANs Pix2Pix
CycleGAN
Everybody dance now
Pix2PixHD
A Style-Based Generator Architecture for Generative Adversarial Networks
BigGAN
Towards a Deeper Understanding of Adversarial Losses
7/4 Self-supervision
Ben Finestein & Ofer Israelov Self Supervision Context Encoders: Feature Learning by Inpainting (al et Darrel, Efros)
Efros, Jigsaw ICCV'15
Jigsaw, Noroozi & Favaro
Learning to count, Noroozi, Pirisivash, Favaro
Colorful Image Colorization, Zhang, Isola, Efros
Tracking Emerges by Colorizing Videos, Vondrick et al
Learning Correspondence from the Cycle-Consistency of Time, Wang, Jabri, Efros
14/4 Understanding CNN properties
Liad Polak & Assaf Shocher CNN props Understanding deep learning requires rethinking generalization
Catastrophic forgetting, Bengio, ICLR19
Effective Receptive Field, Urtasun, NIPS16
Are All Layers Created Equal?, Yoram Singer
21/4 NO CLASS (PASSOVER)
28/4 One/few shot learning
Michael Roitman & Shir Amir one/few shot learning Deepmind, Matching Networks for One Shot Learning
Karlinsky, RepMet
Karlinsky, Delta-encoder
Learning to Compare
A closer look at few-shot classification, accepted ICLR19
5/5 Non-local neural networks
Gil Pinsky & Hila Levi non local Attention is all you need
Non-local Neural Networks, Wang, Girshick, Gupta, He
A simple neural network module for relational reasoning
Relational Deep Reinforcement Learning
Non-Local Recurrent Network for Image Restoration
Feature Denoising for Improving Adversarial Robustness
Self attention GAN, Zhang, Goodfellow Metaxas, Odena
12/5 Transfer/multi-task learning
Lior Yariv & Niv Granot transfer / multitask learning Taskonomy (Zamir, Malik) (best paper CVPR18)
Partial Transfer Learning with Selective Adversarial Networks (Michael Jordan, CVPR18)
Multi-task learning as multi-objective optimization (Koltun, NIPS18)
19/5 Neural architecture search
Dror Moran & Maayan Tamari NAS NAS with RL, ICLR17
DeepMind&CMU, Differentiable NAS
Simple NAS for CNNs, ICLR18
Efficient NAS, ICML18
Image-net SotA
NAS over a graph
Urtasun, accepted ICLR19
Survey
26/5 Image Enhancement & Quality assesment
Gil Karni & Shai Sapir Deep Features as a Perceptual Metric, Isola, Efros, Shechtman, CVPR18
Perception-Distortion, Blau, Michaeli, CVPR18
2/6 Vision and Audio
Ilan Aizelman & Natan Bibelnik lip_reading
audio_visual
Looking to listen at the cocktail party, SIG18
The sound of pixels, Torralba, ECCV18
Self-Supervised, Efros, ECCV18 oral
2 papers, Zisserman, ECCV18
X2Face, Zisserman ECCV18
Learning to lip read words by watching videos
Deep Lip Reading: a comparison of models and an online application
9/6 NO CLASS (SHAVUOT)
16/6 NO CLASS (CVPR conference)
23/6 Semi-supervised learning
Shahaf Wagner & Omri Kramer semi-supervised Towards Omni-Supervised Learning- He, CVPR18
Realistic Evaluation of Deep Semi-Supervised Learning Algorithms- Goodfellow NIPS18
MixMatch: A Holistic Approach to Semi-Supervised Learning (By authors from Google Brain including Ian Goodfellow)
Billion-scale semi-supervised learning for image classification (By Facebook AI)
X2Face, Zisserman ECCV18
Learning to lip read words by watching videos
Deep Lip Reading: a comparison of models and an online application
30/6 Interpretability & Visualization
Shmuel Fine & Amir Goren
7/7 Adversarial attacks and defences
Matan Schlanger & Samuel Londner