Machine Learning Advances Symposium

May 12th, 2023

About MLAS

Machine learning is having a moment right now. It is increasingly becoming an integral part of business, healthcare, finance, and other aspects of our day-to-day lives. Meanwhile, the research at our own CSAIL and LIDS communities is driving many of the major advances in machine learning. This event will bring together the CSAIL and LIDS communities to share our cutting-edge and diverse lines of research, to identify the biggest challenges facing machine learning, and to brainstorm what comes next.
Complete the registration form to attend

Speakers

Dylan Hadfield-Menell

CSAIL

Cathy Wu

LIDS

Yoon Kim

CSAIL

Ashia Wilson

LIDS

Polina Golland

CSAIL

Martin Wainright

LIDS

Event Schedule​

Events will be held in MIT Building E15:
Bartos theater (for talks)
Lower atrium (for posters and breaks)

09:00 AM​

Opening Remarks

09:15 AM​

Talk

Polina Golland

Learning to read x-rays

09:45 AM​

Spotlight Talks

Abhin Shah: On counterfactual inference with unobserved confounding

Alex Gu: ObSynth: An Interactive Synthesis System for Generating Object Models from Natural Language Specifications

10:00 AM​

Coffee Break​

10:30 AM​

Talk

Yoon Kim

Large Language Models & Symbolic Structures

11:00 AM​

Talk

Cathy Wu

Intelligent Vehicle Coordination: Eco-driving for the Planet and Hybrid Learning Methods

11:30 AM​

Spotlight Talks

Tongzhou Wang: Quasimetric Reinforcement Learning

Charles Dawson: Root-cause analysis and repair of failures in autonomous decision-making

Saachi Jain: Distilling Model Failures as Directions in Latent Space

Kwangjun Ahn: Reproducibility in Optimization: Theoretical Framework and Limits

12:00 PM​

Posters & Standup Lunch

List of posters

01:30 PM​

Talk

Martin Wainwright

Beyond worst-case: Instance-optimal algorithms in ML

2:00 PM​

Talk

Ashia Wilson

02:30 PM​

Coffee Break

03:00 PM​

Talk

Dylan Hadfield-Menell

03:30 PM​

Spotlight Talks

Manon Revel: Diversity and Expertise in Binary Votes Aggregation

Aparna Balagopalan: Judging Facts, Judging Norms: Training Machine Learning Models to Judge Humans Requires a Modified Approach to Labeling Data

Renato Berlinghieri: Gaussian processes at the Helm(holtz): A more fluid model for ocean currents

Joanna Materzynska: Erasing Concepts from Diffusion Models

04:00 PM​

Posters & Drinks/Snacks

List of posters

05:30 PM​

Social

Organizers

Chairs
Tamara Broderick, Associate Professor, EECS
Phillip Isola, Associate Professor, EECS

Committee
Aleksander Madry, Professor, EECS
Devavrat Shah, Professor, EECS
Marzyeh Ghassemi, Assistant Professor, EECS
Alexander Rakhlin, Professor, BCS
Suvrit Sra, Associate Professor, EECS
Jacob Andreas, Assistant Professor, EECS

Our Sponsors​