Hauptseminar/Master Seminar:
Deep Generative Models WS2018

Team: Prof. Nassir Navab, Dr. Vasileios Belagiannis, Dr. Federico Tombari, Iro Laina, Magda Paschali, Leslie Casas, Oliver Scheel, Yida Wang, Helisa Dhamo

Type: Advanced Seminar Course Module IN0014
Type: Master Seminar Module IN2107
Type: Seminar Module IN4826
SWS: 2+0
ECTS: 5 Credits
Location: MI 03.13.010
Time: TBD
Course Language: English

Announcements

• 28.06.2018 - Motivation Letters are welcomed.
• 12.06.2018 - Preliminary meeting 21.06.2018 at 5:00pm, MI 03.13.010.

Introduction

• The recent advances of generative models in Deep Learning will be studied in the Seminar Course. Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), adversarial learning and learning with adversaries compose the main topics of the seminar. A set of papers will cover the aforementioned topics. One or two lectures will be also offered to support the seminar.

Registration

• The registration is though the TUM Matching Platform (from 29.06 to 11.07).

Requirements

In this Seminar (formerly Hauptseminar), each student is asked to select one paper from a list. In order to successfully complete the seminar, participants have to fulfill these requirements:

• Presentation: The selected paper is presented to the other participants (20 minutes presentation 10 minutes questions). Use the CAMP templates for PowerPoint camp-tum-jhu-slides.zip, or Latex: CAMP-latex-template.
• Written Report: A document of maximum 8 pages is written and submitted one week after the presentation. Please download and use the llncs2e.zip Latex template.
• Attendance: Participants have to participate actively in all seminar sessions.

The students are required to attend each seminar presentation which will be held during this course. Each presentation is followed by a discussion and everyone is encouraged to actively participate. Part of the discussions outcome will be included in the report too. The report must include all references used and must be written completely in your own words. Copy and paste will not be tolerated. Both report and presentation have to be done in English.

List of Publications

Nr	Authors	Title	Conference	Link	Tutor	Student
01	Anish Athalye et. al.	Synthesizing Robust Adversarial Examples	ICML 2018	https://arxiv.org/abs/1707.07397	-	-
02	Anish Athalye et. al.	Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples	ICML 2018	https://arxiv.org/abs/1802.00420		-
03	Andrew Ilyas et. al.	Black-box Adversarial Attacks with Limited Queries and Information	ICML 2018	https://arxiv.org/abs/1804.08598	Oliver	Paolo N.
04	Lars Mescheder et. al.	Which Training Methods for GANs do actually Converge?	ICML 2018	https://arxiv.org/abs/1801.04406	Azade	Ali Abbas J.
05	Xiaoyu Lu et. al.	Structured Variationally Auto-encoded Optimization	ICML 2018	http://proceedings.mlr.press/v80/lu18c.html	Leslie	Julio O.
06	Junbo Zhao et. al.	Adversarially Regularized Autoencoders	arxiv	https://arxiv.org/abs/1706.04223	-	-
07	Jiawei Su et. al.	One pixel attack for fooling deep neural networks	arxiv	https://arxiv.org/abs/1710.08864	Magda	Jongwon L.
08	Karol Kurach et. al.	The GAN Landscape: Losses, Architectures, Regularization, and Normalization	arxiv	https://arxiv.org/abs/1807.04720	Helisa	Ivan P.
09	Mario Lucic, et. al.	Are GANs Created Equal? A Large-Scale Study	arxiv	https://arxiv.org/abs/1711.10337	Vasilis	Maria M.
10	Angjoo Kanazawa, et. al.	Learning Category-Specific Mesh Reconstruction from Image Collections	ECCV 2018	https://arxiv.org/abs/1803.07549	Yida	Giorgio F.
11	Omid Poursaeed, et. al.	Generative Adversarial Perturbations	CVPR 2018	https://arxiv.org/abs/1712.02328	Leslie	Abdelrahman E.
12	Samaneh Azadi, et. al.	Multi-Content GAN for Few-Shot Font Style Transfer	CVPR 2018	https://arxiv.org/abs/1712.00516	Yida	Wout B.
13	Ming-Yu Liu, et. al.	Unsupervised Image-to-Image Translation Networks	NIPS 2017	https://arxiv.org/abs/1703.00848	Federico	Aadhithya S.
14	Long Chen, et. al.	Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Networks	CVPR 2018	https://arxiv.org/abs/1712.01928	Azade	Mingyip C.
15	Xi Peng, et. al.	Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation	CVPR 2018	https://arxiv.org/abs/1805.09707	Iro	Hanna K.
16	Minghua Zhang, et. al.	An Unsupervised Model with Attention Autoencoders for Question Retrieval	AAAI 2018	https://arxiv.org/abs/1803.03476	-	-
17	Chaowei Xiao, et. al.	Spatially Transformed Adversarial Examples	ICLR 2018	https://arxiv.org/abs/1801.02612	Magda	Nitin V.
18	Sashank J. Reddi, et. al.	ON THE CONVERGENCE OF ADAM AND BEYOND	ICLR 2018	https://openreview.net/forum?id=ryQu7f-RZ	-	-
19	Jonathan Ho and Stefano Ermon	Generative Adversarial Imitation Learning	NIPS 2016	https://arxiv.org/abs/1606.03476	Vasilis	Raymond C.

Schedule

The schedule for the paper presentations will be announced here during the semester. Each student is required to attend all presentations.

Date	Time	Title	Speaker	Material
17.10.2018	16:00 - 18:00	Lecture, Are GANs Created Equal? A Large-Scale Study, The GAN Landscape: Losses, Architectures, Regularization, and Normalization.	Maria M., Ivan P.	Slides
08.11.2018	16:00 - 18:00	One pixel attack for fooling deep neural networks, Spatially Transformed Adversarial Examples, Black-box Adversarial Attacks with Limited Queries and Information, Zero-Shot Visual Recognition using Semantics-Preserving Adversarial Embedding Networks,	Jongwon L., Nitin V., Paolo N. Mingyip C.
15.11.2018	16:00 - 18:00	Generative Adversarial Perturbations, Generative Adversarial Imitation Learning, Which Training Methods for GANs do actually Converge?, Multi-Content GAN for Few-Shot Font Style Transfer	Abdelrahman E., Raymond C., Ali Abbas J., Wout B.
22.11.2018	16:00 - 18:00	Unsupervised Image-to-Image Translation Networks, Learning Category-Specific Mesh Reconstruction from Image Collections, Structured Variationally Auto-encoded Optimisation Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation	Aadhithya S., Giorgio F., Julio O., Hanna K.