ICON Seminar in Learning: Prof. David I. Inouye (Purdue ECE)
| Event Date: | October 21, 2022 |
|---|---|
| Speaker: | Prof. David I. Inouye |
| Speaker Affiliation: | Purdue University |
| Time: | 2-2:30 pm |
| Location: | HAMP 1252 (in-person)
Zoom: https://purdue-edu.zoom.us/j/99326083199?pwd=TjhqaE1TYkZ2RG1RMkt6S1JZbGk1UT09 |
| Contact Name: | Yu She |
| Contact Email: | yushe@purdue.edu |
| Priority: | No |
| School or Program: | College of Engineering |
| College Calendar: | Show |
Location: HAMP 1252 (in-person) and Zoom (online) https://purdue-edu.zoom.us/j/99326083199?pwd=TjhqaE1TYkZ2RG1RMkt6S1JZbGk1UT09
Agenda: 2:00pm-3:15pm (Seminar + Q&A); 3:15pm-4:00pm (Networking). Snacks and coffee will be provided.
Unifying and Advancing the Science of Deep Distribution Alignment
Abstract
Distribution alignment has the opposite objective of classification. While classification finds a representation that separates two distributions, alignment finds a representation that brings together two distributions. Alignment has been used to enhance robustness or enforce constraints in many recent machine learning applications including domain generalization, causal discovery, and fair representation learning. Despite these important applications, distribution alignment research lacks a unified and systematic conceptual framework and has primarily focused on GAN-based adversarial alignment for images. To address this gap, I will present a unifying alignment framework that encompasses alignment concepts, measures, algorithms, and applications. Specifically, I will formalize the definition of distribution alignment, develop novel non-adversarial alignment measures and algorithms, and discuss alignment applications in causal discovery and domain generalization. Ultimately, this work aims to advance the science of distribution alignment to enable the next generation of contextually aware and robust AI systems.
Bio
Prof. David I. Inouye is an assistant professor in the Elmore Family School of Electrical and Computer Engineering at Purdue University. He leads the Probabilistic and Understandable Machine Learning Lab, which focuses on the fundamentals of distribution alignment, probabilistic models, and explainable AI. More recently, he is interested in distribution alignment including new alignment algorithms, measures, and applications such as causality and domain generalization. On the explainable AI side, he is interested in distribution shift explanations and tractable uncertainty quantification. Previously, he was a postdoc at Carnegie Mellon University working with Prof. Pradeep Ravikumar. He completed his Computer Science PhD at The University of Texas at Austin in 2017 advised by Prof. Inderjit Dhillon and Prof. Pradeep Ravikumar. He was awarded the NSF Graduate Research Fellowship (NSF GRFP).
Seminar Photoes
Seminar Recording:
https://youtu.be/2jWLFwewc_M
2022-10-21 14:00:00 2022-10-21 14:30:00 America/Indiana/Indianapolis ICON Seminar in Learning: Prof. David I. Inouye (Purdue ECE) Title: Unifying and Advancing the Science of Deep Distribution Alignment HAMP 1252 (in-person) Zoom: https://purdue-edu.zoom.us/j/99326083199?pwd=TjhqaE1TYkZ2RG1RMkt6S1JZbGk1UT09