Program Overview

Program Overview

The Rutgers University Department of Physics and Astronomy offers a course titled "Modern Machine Learning for Physics and Astronomy" (PHYSICS 693, FALL 2023). This course surveys state-of-the-art techniques in modern machine learning and their applications to physics and astronomy.

Course Description

Deep learning and neural networks are transforming data analysis in nearly every branch of physics and astronomy, leading to new discoveries, more sensitive measurements, and faster and more powerful simulations. The course will cover machine learning methods, including classification/regression, generative modeling, and anomaly detection. Students will be introduced to major ML frameworks and architectures such as CNNs, transformers, GANs, VAEs, normalizing flows, and diffusion models.

Applications and Subfields

These methods will be illustrated through their applications to a broad array of subfields, including:

• Particle physics
• Astronomy
• Cosmology
• Condensed matter physics
• More

Prerequisites and Requirements

• No prior knowledge of modern ML or domain knowledge of any of these fields is required.
• However, prior experience in data analysis and coding (Python, numpy, matplotlib, ...) would be helpful to get the most out of the course.

Course Objectives

Along the way, students will receive a practical introduction to fundamental concepts and methods in machine learning and data science. The course aims to enable students to understand current research trends in ML applications to physics and astronomy and embark on research projects of their own.

General Information

Instructor

• Prof. David Shih

Lectures

• MTh 10:20-11:40, NHETC SEMINAR ROOM

Resources

• The Deep Learning Book. Goodfellow, Bengio, and Courville (2016).

Assessment and Evaluation

Homeworks

• TBD

Exams

• TBD

Lecture Notes

The course includes a series of lectures, with topics and dates specified:

• Lecture 1: Thursday, September 7, 2023
• Lecture 2: Monday, September 11, 2023
• Lecture 3: Thursday, September 14, 2023
• Lecture 4: Monday, September 18, 2023
• Lecture 5: Thursday, September 21, 2023
• Lecture 6: Monday, September 25, 2023
• Lecture 7: Thursday, September 28, 2023
• Lecture 8: Monday, October 2, 2023
• Lecture 8: Monday, October 9, 2023
• Lecture 9: Thursday, October 12, 2023
• Class presentations: Monday, October 16, 2023
• Lecture 10: Thursday, October 19, 2023
• Lecture 11: Monday, October 23, 2023
• Lecture 12: Thursday, October 26, 2023
• Lecture 13: Monday, October 30, 2023
• Lecture 14: Thursday, November 2, 2023
• Lecture 15: Thursday, November 9, 2023 (SPECIAL GUEST LECTURE: DARIUS FAROUGHY)
• Lecture 16: Monday, November 13, 2023
• Lecture 17: Thursday, November 16, 2023 (SPECIAL GUEST LECTURE: DARIUS FAROUGHY)
• Lecture 18: Monday, November 20, 2023
• Lecture 19: Tuesday, November 21, 2023
• Lecture 20: Monday, November 27, 2023
• Class presentations: Thursday, November 30, 2023
• Lecture 21: Monday, December 4, 2023
• Lecture 22: Monday, November 11, 2023

Demos

The course includes demos on various topics, including:

• 10d Gaussian toy model -- DNN binary classifier and Neyman-Pearson comparison (Lecture 5)
• Top tagging -- cut-based and DNN classifier with high-level features (Lecture 5)
• MNIST generation -- vanilla GAN (Lecture 11)
• GAN Lab demo (illustrates mode collapse) (Lecture 12)
• MNIST generation -- WGAN-GP (Lecture 13)
• MNIST generation and latent space visualization -- VAE (Lecture 14)