Classes start on January 16, 2018
Date | Topic | Reading | Notes |
---|---|---|---|
Tue, Jan 16 | Course Overview |
CMU Computing Policy CMU Policy on Academic Integrity Nature Review on Deep Learning Classification Intro |
Course overview Note |
PART I: Machine Learning Basics | |||
Thur, Jan 18 | Intro to Machine Learning I (Linear Algebra Review, Logistic Regression) |
Deep Learning Book, CH2 Linear Classification Basics |
Notes |
Tue, Jan 23 | Intro to Machine Learning II (Gradient Descent, Stochastic Gradient Descent)) |
Stochastic Gradient Descent Deep Learning Book, CH3 |
Notes |
Thur, Jan 25 | Intro to Machine Learning III (Stochastic Gradient Descent)) HW1Out |
Stochastic Gradient Descent Deep Learning Book, CH4 + CH8:8.1.3 |
Notes |
PART II: Intro to Deep Learning | |||
Tue, Jan 30 | Deep Learning Software (In preparation for HW 1) |
Documentation of Keras Documentation of Theano |
Notes |
Thur, Feb 1 | Back Propagation |
Back Propagation |
Notes |
Tue, Feb 6 | Deep Feedforward Networks |
Deep Learning Book, CH6 Setting up the Data and Model |
Notes |
Thur, Feb 8 | DFN Training |
Deep Learning Book, CH6 Learning and Evaluation |
Notes |
Tue, Feb 13 | Convolutional Neural Networks I |
CNN Architecture Deep Learning Book, CH9 |
Notes |
Thur, Feb 15 | Convolutional Neural Networks II HW1 In |
CNN Visualization Deep Learning Book, CH9 |
Notes |
PART III: Explanation for Deep Neural Networks | |||
Tue, Feb 20 | Paper Discussion: Axiomatic Attribution for Deep Networks Guest Lecture Ankur Taly (Google) |
Axiomatic Attribution for Deep Networks | Notes |
Thur, Feb 22 | Deep Learning Software (In preparation for HW 2) HW2 Out : code files |
Documentation of Keras Documentation of Tensorflow |
Notes |
Tue, Feb 27 | Paper Discussion: Influence-Directed Explanations for CNN | Influence-Directed Explanations for CNN | Notes |
Thur, Mar 1 |
Paper Discussion: Influence functions Guest Lecture: Pang Wei Koh (Stanford) |
Understanding Black-box Predictions via Influence Functions | Notes |
Tue, Mar 6 | Paper Discussion: Influence-Directed Explanations for CNN (Continued) | Influence-Directed Explanations for CNN | Notes |
PART IV: Adversarial Learning | |||
Thur, Mar 8 | Paper Discussion: Adversarial Settings in Deep Learning Guest Lecture: Matt Fredrikson(CMU) |
The Limitations of DL in Adversarial Settings | notes |
Tue, Mar 13 | No Class: Spring Break | ||
Thur, Mar 15 | No Class: Spring Break | ||
Tue, Mar 20 | Paper Discussion: Defensive Distillation Guest Lecture: Nicholas Carlini (UC Berkeley) HW3 In |
Towards Evaluating the Robustness of Neural Networks | notes |
Thur, Mar 22 | Paper Discussion: Detecting Adversarial Samples Guest Lecture: Saurabh Shintre (Symantec) HW3 Out |
Detecting Adversarial Samples from Artifacts | notes |
Tue, Mar 27 | HW3 Preparation | TBD | notes |
Thur, Mar 29 | Paper Discussion: Generative Adversarial Networks | Generative Adversarial Nets | notes |
PART V: Recurrent Neural Networks | |||
Tue, Apr 3 | Recurrent Neural Network I | Deep Learning Book, CH10 | notes |
Thur, Apr 5 | Recurrent Neural Network II | Deep Learning Book, CH10 | notes |
Tue, Apr 10 | Recurrent Neural Network III | Deep Learning Book, CH10 | notes |
Thur,Apr 12 | Paper Discussion: Bias In Word Embeddings Guest Lecture: James Zou(Stanford) |
Man is to Computer Programmer as Woman is to Homemaker? Human-like Bias in Langauge Models |
notes |
Tue, Apr 17 | Paper Discussion: Explanation for RNN I | A causal framework for explaining the predictions of black-box sequence-to-sequence models | notes |
Thur, Apr 19 | Paper Discussion: Explanation for RNN II HW3 In HW4 Out |
Rationalizing Neural Predictions | notes |
Tue, Apr 24 | Explanation for RNN III | TBD | notes |
Thur, Apr 26 | TBD | TBD | notes |
Tue, May 1 | TBD | TBD | notes |
Thur, May 3 | Class Wrapup | TBD | notes |
* All schedules are subject to change over the course.