CS839 Fall 2020

Tentative Schedule (Subject to Change)

Week	Date	Topic	Reading materials	Assignments
1	Thursday, September 3	Course overview and introduction [Slides]		Sign up for paper presentations and scribes (link to Google sheet)
2	Tuesday, September 8	Neural Architecture Design I (lecture) [Slides]	Goodfellow-Bengio-Courville Chapter 6
2	Thursday, September 10	Neural Architecture Design II [Slides][Notes]	ImageNet Classification with Deep Convolutional Neural Networks Deep Residual Learning for Image Recognition Densely Connected Convolutional Networks (to present) Mask R-CNN
3	Tuesday, September 15	Neural Architecture Design III [Slides][Notes]	Neural Architecture Search with Reinforcement Learning Efficient Neural Architecture Search via Parameter Sharing (to present)
3	Thursday, September 17	Trustworthy Deep Learning I (lecture) [Slides]	Goodfellow-Bengio-Courville Chapter 7.5, 7.13	Team registration (link to Google sheet)
4	Tuesday, September 22	Trustworthy Deep Learning II [Slides][Notes]	Out-of-distribution Reliability Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images On Calibration of Modern Neural Networks Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Data Informative Outlier Matters: Robustifying Out-of-distribution with Outlier Mining (to present) Energy-based Out-of-distribution Detection
4	Thursday, September 24	Trustworthy Deep Learning III [Slides][Notes]	Adversarial Robustness Towards Deep Learning Models Resistant to Adversarial Attacks Intriguing properties of neural networks Robust Physical-World Attacks on Deep Learning Models Unlabeled Data Improves Adversarial Robustness Defense Against Adversarial Images using Web-Scale Nearest-Neighbor Search Understanding and Mitigating the Tradeoff Between Robustness and Accuracy (to present)
5	Tuesday, September 29	Trustworthy Deep Learning IV [Slides][Notes]	Fairness / Group Robustness An investigation of why overparameterization exacerbates spurious correlations (to present) Invariant Risk Minimization Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization Model Patching: Closing the Subgroup Performance Gap with Data Augmentation
5	Thursday, October 1	No class (use the time to discuss the project proposal with team)		Project proposal submission deadline on October 4 Midnight (download proposal latex template here)
6	Tuesday, October 6	Meetings with Instructor to Review Project Proposals		Signup sheet for 15-min slot
8	Thursday, October 8	Interpretable Deep Learning I (Lecture) [Slides]	Guest lecture by TA: Yiyou Sun
9	Tuesday, October 13	Interpretable Deep Learning II [Slides][Notes]	Attribution Methods Learning Deep Features for Discriminative Localization (CAM) Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization (to present) Network Dissection: Quantifying Interpretability of Deep Visual Representations Axiomatic Attribution for Deep Networks
9	Thursday, October 15	Interpretable Deep Learning III [Slides][Notes]	Rethinking interpretability Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead (to present) Interpretation of Neural Networks is Fragile The (Un)Reliability of Saliency Methods
10	Tuesday, October 20	Deep Learning Generalization and Theory I (Lecture) [Slides]	Guest speaker: Yuandong Tian (Facebook AI Research)
10	Thursday, October22	Deep Learning Generalization and Theory II [Slides][Notes]	Exploring Generalization in Deep Learning Towards Understanding the Role of Over-Parametrization in Generalization of Neural Networks (to present) On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima Understanding deep learning requires rethinking generalization Visualizing the Loss Landscape of Neural Nets
11	Tuesday, October 27	Deep Learning Generalization and Theory III [Slides][Notes]	One ticket to win them all: generalizing lottery ticket initializations across datasets and optimizers The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks (to present)
6	Thursday, October 29	Learning with Less Supervision I (Lecture) [Slides]
7	Tuesday, Novermber 3	Learning with Less Supervision II [Slides][Notes]	Weakly supervised learning Big Transfer (BiT): General Visual Representation Learning (to present) Exploring the Limits of Weakly Supervised Pretraining Learning Visual Features from Large Weakly Supervised Data Revisiting Unreasonable Effectiveness of Data in Deep Learning Era Snorkel: Rapid Training Data Creation with Weak Supervision
7	Thursday, November 5	Learning with Less Supervision III [Slides][Notes]	Unsupervised learning SCAN: Learning to Classify Images without Labels A Simple Framework for Contrastive Learning of Visual Representations (to present) Self-Supervised Learning of Pretext-Invariant Representations Rethinking Pre-training and Self-training
8	Tuesday, November 10	Learning with Less Supervision IV [Slides][Notes]	Data Augmentation (some blogs I wrote) Automating the Art of Data Augmentation: Overview Automating the Art of Data Augmentation: Practice Automating the Art of Data Augmentation: Theory Automating the Art of Data Augmentation: New Direction Unsupervised Data Augmentation for Consistency Training (to present)
11	Thursday, November 12	Continual / Lifelong Learning I (Lecture) [Slides]
12	Tuesday, November 17	Continual / Lifelong Learning II [Slides][Notes]	Overcoming Catastrophic Forgetting with Unlabeled Data in the Wild (to present) iCaRL - Incremental Classifier and Representation Learning LwF-Learning without Forgetting Overcoming catastrophic forgetting in neural networks
12	Thursday, November 19	Continual / Lifelong Learning III [Slides][Notes]	SupSup - Supermasks in Superposition (to present) PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning Progressive Neural Networks
13	Tuesday, November 24	Deep Generative Model I (Lecture) [Notes on GAN][Notes on style transfer GAN]	Goodfellow-Bengio-Courville Chapter 20
13	Thursday, November 26	No class (Thanksgiving)
14	Tuesday, December 1	Deep Generative Model II [Slides][Notes]	Generative Adversarial Networks InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets Improved Techniques for Training GANs Stacked Generative Adversarial Networks Large Scale GAN Training for High Fidelity Natural Image Synthesis Wasserstein GAN (to present)
14	Tuesday, December 3	Deep Generative Model III	Style Transfer Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks A Neural Algorithm of Artistic Style (to present) A Style-Based Generator Architecture for Generative Adversarial Networks
15	Tuesday, December 8	Final project presentation (Team 1-9)
15	Tuesday, December 10	Final project presentation (Team 10-17)
16	Thursday, December 17	Final project written report due (by end of the day)

Special Topics in AI: Deep Learning

Tentative Schedule (Subject to Change)