Deep Learning

Columbia University - Fall 2019

Class is held in 451 CS on Mon,Wed 6:40-7:55pm

Monday 4:30-5:30pm, CSB 453: Lecturer, Iddo Drori

Tuesday 11am-12pm, TA room: Course Assistant, Samrat Phatale

Wednesday 11am-12pm, TA room: Course Assistant, Benedikt Schifferer

Thursday 9-10am, TA room: Course Assistant, Xiren Zhou

Friday 11am-12pm: Course Assistant, Abhay Khosla

CVN Course Assistant, Manik Goyal

First Day of Classes (Tuesday, September 3)

Lecture 1 (Wednesday, September 4): Introduction
Machine learning timeline: from Least Squares to AlphaZero, Deep CFR, and BERT, milestones of neural networks and deep learning. Linear algebra review, fully connected neural networks, forward propagation as a composition of functions, each with linear and non-linear component, nonlinear activation functions, network loss functions.

Lecture 2 (Monday, September 9): Backpropagation

Lecture 3 (Wednesday, September 11): Optimization
Tutorial 1: TensorFlow

Exercise 1 (Friday, September 13): due Friday, Sep. 20

Lecture 4 (Monday, September 16): CNNs
Tutorial 2: Keras

Lecture 5 (Wednesday, September 18): CNNs
Tutorial 3: PyTorch

Exercise 2 (Friday, September 20): due Friday, Sep. 27

Lecture 6 (Monday, September 23): Regularization, sequence models

Lecture 7 (Wednesday, September 25): Sequence models

Competition (Friday, September 27 - Monday October 21): Learning to drive

Lecture 8 (Monday, September 30): Sequence models
Tutorial 4: AllenNLP

Lecture 9 (Wednesday, October 2): Generative models

Lecture 10 (Monday, October 7): Generative models

Lecture 11 (Wednesday, October 9): Generative models

Lecture 12 (Monday, October 14): Graph neural networks

Lecture 13 (Wednesday, October 16): Reinforcement learning

Lecture 14 (Monday, October 21): Winning the ICCV 2019 Learning to Drive Challenge🏆🚘

Lecture 15 (Wednesday, October 23): Reinforcement learning

Lecture 16 (Monday, October 28): Deep reinforcement learning

Lecture 17 (Wednesday, October 30): Deep reinforcement learning

Academic Holiday (Monday, November 4)

Election Day, University Holiday (Tuesday, November 5)

Lecture 18 (Wednesday, November 6): Deep reinforcement learning

Lecture 19 (Monday, November 11): Imperfect information games, Deep CFR, OpenSpiel

Lecture 20 (Wednesday, November 13): Automated deep learning

Lecture 21 (Monday, November 18): Deep learning systems, recommender systems, fairness

Lecture 22 (Wednesday, November 20): Deep learning applications

Lecture 23 (Monday, November 25): Deep learning in computer vision

Academic Holiday (Wednesday, November 27)

Thanksgiving Day, University Holiday (Thursday, Nov. 28)

Academic Holiday (Friday, November 29)

Lecture 24 (Monday, December 2): Quantum computing (IBM QX, AWS Braket)

Lecture 25 (Wednesday, December 4): Quantum neural networks

Lecture 26 (Monday, December 9): Project poster sessions
Beauty GAN vs. Beauty Glow: A neural network judge
Art synthesis using SPADE
Art synthesis using style GAN
Movie trailer poster synthesis using GAN's
Text to image synthesis using style-based attention GAN
Text guided human pose synthesis
Zero-shot sentiment analysis
Human animation by semantic parsing and pose keypoints
Cartoon interpolation
3D pose and style synthesis
Deep view synthesis using an octree representation
Adversarial robustness in audio
Imperceptible voice swapping from a few utterances
Voice swapping
Voice cloning
Bayesian optimization for deep CFR
Stock trading using deep reinforcement learning and optimization
Foreign exchange trading by deep reinforcement learning
End-to-end learning for stock trading
Reinforcement learning and sequence models for stock trading
Financial time series modeling by deep reinforcement learning
Deep reinforcement learning for stock trading x 3
Can SATNet learn crossword puzzles?
VRP approximation using reinforcement learning
TSP approximation using GNN's
E-hailing driver repositioning by multi-agent reinforcement learning
Metalloprotein design using a CVAE
Learning fair ranking policies
Extracting data from bar charts using CNN's
Lyft 3D object detection for autonomous vehicles competition x 3

Last Day of Classes (Monday, December 9)