Deep Learning

Boston University - Summer I 2023

Class is held in HAR 212 on Tuesday and Thursday 2-5:30pm

Course staff and office hours

Instructor: Prof. Iddo Drori, TBD

Teaching Fellow, TBD

Grader, TBD

Textbook

The Science of Deep Learning, Iddo Drori, Cambridge University Press, 2022

Enrolled students receive a free online version

Part I: Foundations

Lecture 1 (Tuesday, May 23): Introduction

Lecture 2 (Tuesday, May 23): Forward and Backpropagation

Lecture 3 (Thursday, May 25): Optimization

Lecture 4 (Thursday, May 25): Regularization

Part II: Architectures

Lecture 5 (Tuesday, May 30): Convolutional neural networks (CNNs)

Lecture 6 (Tuesday, May 30): Sequence models (RNNs, LSTM, GRU)

Lecture 7 (Thursday, June 1): Graph neural networks (GNNs)

Lecture 8 (Thursday, June 1): Transformers

Part III: Foundation Models

Lecture 9 (Tuesday, June 6): GPTs

Lecture 10 (Tuesday, June 6): GPTs

Lecture 11 (Thursday, June 8): Vision, audio, and language

Lecture 12 (Thursday, June 8): Vision, audio, and language

Part IV: Generative Models

Lecture 13 (Tuesday, June 13): Diffusion models

Lecture 14 (Tuesday, June 13): Generative adversarial networks (GANs)

Lecture 15 (Thursday, June 15): Variational autoencoders (VAEs)

Lecture 16 (Thursday, June 15): Language and vision

Part V: Reinforcement Learning

Lecture 17 (Tuesday, June 20): Markov decision processes

Lecture 18 (Tuesday, June 20): Reinforcement learning

Lecture 19 (Thursday, June 22): Deep reinforcement learning

Lecture 20 (Thursday, June 22): Deep reinforcement learning

Part VI: Applications

Lecture 21 (Tuesday, June 27): Deep learning for education

Lecture 22 (Tuesday, June 27): Deep learning for climate science

Lecture 23 (Thursday, June 29): TBD

Lecture 24 (Thursday, June 29): TBD

Exercises: quiz and programming homework

Exercise 1: Forward and Backpropagation, Optimization

Exercise 2: CNN's, RNN's

Exercise 3: GNN's, Transformers

Exercise 4: DDPM's, VAE's

Exercise 5: RL

Tutorials

Tutorial 1: PyTorch

Tutorial 2: Keras

Tutorial 3: CNN's

Tutorial 4: RNN's

Tutorial 5: dgl.ai, GNN library

Tutorial 6: huggingface.co, Transformers library

Tutorial 7: RLlib, reinforcement learning library