ETH Zurich - D-INFK - IVC - CVG - Lectures - Deep Learning Seminar

Deep Learning Seminar


Course title: Deep Learning for Computer Vision: Seminal Work
Course ID: 263-5904-00L
Lecturers: Dr. Zhaopeng Cui, Dr. Martin Oswald
Teaching assistants: Ian Cherabier, Mihai Dusmanu, Marcel Geppert, Viktor Larsson, Zuoyue Li, Sandro Lombardi, Daniel Thul
Venue: Mo 15-17h in CAB G 57

This seminar covers seminal papers on the topic of deep learning for computer vision. The students will present and discuss the papers and gain an understanding of the most influential research in this area - both past and present. The objectives of this seminar are two-fold. Firstly, the aim is to provide a solid understanding of key contributions to the field of deep learning for vision (including a historical perspective as well as recent work). Secondly, the students will learn to critically read and analyse original research papers and judge their impact, as well as how to give a scientific presentation and lead a discussion on their topic.

Each student chooses one paper from the provided collection to present during the course of the seminar. The students will be supported in the preparation of their presentation by the seminar assistants.

News

  • 15.04. Paper #22 (AtlasNet) moved from 20.05. to 29.04. There will be no seminar on the 20.05.
  • 04.03. Please upload the slides in the Moodle course after your presentation.
  • 04.03. Paper assignment has been updated: the presentation for paper #21 will be cancelled on May 20th.
  • 20.02. Paper assignment has been released.
  • 18.02. Paper assignment will open today at 3pm. Please choose your favorite papers in the Moodle course until the end of tomorrow (19.02.).
  • 18.02. The slides of today's introduction session are online now.

Important information

Each student will present one paper. Everybody is encouraged to read each paper before it is being presented and engage in a discussion following the presentations. To foster interesting discussions, each paper will also be assigned two "critics" who study the paper and prepare questions for the discussion. Each student will be graded based on both their presentation (80%) and their participation in the assigned discussions (20%). Attendance is required to pass the course (3 absences allowed).

Slides

Schedule (Tentative)

Date Topic Presenter Critics Assistant
18.02. Introduction
25.02. 01 ImageNet: A Large-Scale Hierarchical Image Database F. Hasler S. Kellenberger, S. Singh D. Thul
25.02. 02 Playing for Data: Ground Truth from Computer Games B. Walser N. Baumann, C. Fennan M. Oswald
04.03. 03 Learning representations by back-propagating errors M. Mihajlovic F. Hasler, S. Kellenberger I. Cherabier
04.03. 04 Backpropagation applied to handwritten zip code recognition S. Singh C. Fennan, L. Fernandes D. Thul
11.03. 05 ImageNet Classification with Deep Convolutional Neural Networks N. Baumann B. Walser, C. Sprecher M. Oswald
11.03. 06 Going Deeper with Convolutions H. Ho S. Panighetti, F. Hasler D. Thul
18.03. 07 Deep residual learning for image recognition D. Dimitrov M. Mihajlovic, L. Steiner Z. Cui
18.03. 08 Xception: Deep Learning with Depthwise Separable Convolutions C. Sprecher M. Vora, S. Panighetti Z. Cui
25.03. 09 Dropout: a simple way to prevent neural networks from overfitting R. Zenkl H. Ho, B. Walser, M. Tom Z. Cui
25.03. 10 Group Normalization M. Vora N. Storni, M. Flowers Z. Cui
01.04. 11 Visualizing and Understanding Convolutional Networks S. Huang L. Steiner, R. Zenkl M. Dusmanu
01.04. 12 What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? R. Suri D. Jin, M. Mihajlovic Z. Li
15.04. 13 Human-level control through deep re-inforcement learning S. Kellenberger R. Zenkl, D. Dimitrov M. Oswald
15.04. 14 Conditional Random Fields as Recurrent Neural Networks L. Steiner C. Yao, D. Jin M. Oswald
29.04. 15 Mask R-CNN N. Storni R. Suri, H. Ho, S. Huang Z. Cui
29.04. 22 AtlasNet: A Papier-Mâché Approach to Learning 3D Surface Generation D. Paschalidou L. Fernandes, N. Baumann S. Lombardi
06.05. 17 DSAC-differentiable RANSAC for Camera Localization L. Fernandes S. Singh, C. Yao V. Larsson
06.05. 18 Deep Fundamental Matrix Estimation M. Tom D. Dimitrov, D. Paschalidou V. Larsson
13.05. 19 Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks D. Jin M. Flowers, S. Huang Z. Cui
13.05. 20 DeepTAM: Deep Tracking and Mapping S. Panighetti C. Sprecher, R. Suri M. Geppert
27.05. 23 Working hard to know your neighbor's margins: Local descriptor learning loss M. Flowers M. Tom, N. Storni M. Dusmanu
27.05. 24 Dynamic Routing Between Capsules C. Yao D. Paschalidou, M. Vora M. Oswald

© CVG, ETH Zürich lm@inf.ethz.ch