Probabilistic Deep Generative Models Workshop

This repository contains the materials for the Probabilistic Deep Generative Models Workshop held at IIT Madras, July 3-5 2024.

Overview

This workshop aims to introduce participants to Deep and Tractable Probabilistic Generative Models, a special class of generative models that balance expressiveness and tractability. Participants will learn about their theoretical foundations, practical implementations, and real-world applications.

Workshop Schedule

Date: July 3-5, 2024

Organizers:

• Robert Bosch Centre for Data Science and Artificial Intelligence (RBCDSAI)
• Wadhwani School of Data Science & AI (WSAI)
• IIT Madras

Presenters:

• Sriraam Natarajan
• Sahil Sidheekh

Day 1: Foundations of Probabilistic Generative Models

• Introduction to probabilistic models, their importance and applications
• Overview of traditional probabilistic graphical models
• Inference and Learning in graphical models

Day 2: Probabilistic Circuits as Expressive and Tractable Generative Models

• Popular generative models & their tractability for probabilistic inference
• Probabilistic Circuits: Representation and Inference
• Learning Probabilistic Circuits
• Hands on Demo

Day 3: Advanced Topics and Applications

• Deep Probabilistic Circuits
• Hands on Demo
• Applications of Deep Tractable Probabilistic Models

Resources

• Slides
• Hands-On Code Demos
  • Day 2: Building Simple Probabilistic Circuicts with SPFlow
  • Day 2: Einsum Networks: Deep Tensorized PCs in Pytorch
  • Day 3: Implementing a simple Normalizing Flow
  • Day 3: Demo of Probabilistic Flow Circuits

Contact

For any queries, please contact the organizers:

• Robert Bosch Centre for Data Science and Artificial Intelligence (RBCDSAI), IIT Madras
  • Email: rbcdsai@iitm.ac.in
  • Website: RBCDSAI