Date & Time:
November 11, 2024 10:00 am – 11:00 am
Location:
Crerar 346, 5730 S. Ellis Ave., Chicago, IL,
11/11/2024 10:00 AM 11/11/2024 11:00 AM America/Chicago Megha Srivastava (Stanford)- New Challenges of Trust with Large-Scale AI Systems Crerar 346, 5730 S. Ellis Ave., Chicago, IL,

Abstract: Today’s large-scale AI systems, trained with > 200 billion parameters over massive datasets, create new challenges of trust as users have increasingly less control over all aspects of model development.
I will first do a deep dive on the challenge of auditing model training service providers, who currently fine-tune models on behalf of resource-poor clients for a fee without any guarantee of correctness. I will show how prior solutions to this “verifiable training” problem are non-robust due to hardware non-determinism, which we address via a threshold-based rounding scheme over intermediate computations during model training. I will then show how users of modern code-generation models may introduce accidental security vulnerabilities due to misplaced confidence. Finally, I will conclude by discussing ongoing work on the current limitations of methods that seek to establish trust via content provenance (e.g. watermarking, C2PA).

Speakers

Megha Srivastava

PhD Student, Stanford University

Megha Srivastava is a Ph.D. student at Stanford University, co-advised by Dorsa Sadigh and Dan Boneh. She is interested in addressing issues of reliability in machine learning models within the broader context of human-AI interaction. In addition to being supported by the NSF GRFP and IBM Ph.D. Fellowships, her research has been recognized with an ICML Best Paper Runner-Up Award and she was selected as a Rising Star in Machine Learning in 2023.

Related News & Events

UChicago CS News

Saturdays with CSIL — How Undergraduates are Transforming CS Education for Local High School Students

Dec 05, 2024
UChicago CS News

UChicago Researchers Receive Google Privacy Faculty Award for Research on AI Privacy Risks

Nov 22, 2024
UChicago CS News

The Climate App Designed to Tackle Chatham’s Flooding Crisis

Nov 21, 2024
In the News

Globus Receives Multiple Honors in 2024 HPCwire Readers’ and Editors’ Choice Awards

Nov 20, 2024
In the News

Argonne Team Breaks New Ground in AI-Driven Protein Design

Nov 15, 2024
UChicago CS News

DOE Awards Fred Chong and his National Research Team $7.5M to Develop a SMART Software Stack to Control Quantum Computer Noise

Nov 12, 2024
UChicago CS News

CS/LSSG Showcases Sustainability Research and Education

Nov 11, 2024
UChicago CS News

Ph.D. Student Jibang Wu Receives the Stigler Center Ph.D. Dissertation Award for His Work Modeling the Incentive Structures of Reward and Recommendation–Based Systems

Oct 24, 2024
UChicago CS News

Rebecca Willett Receives the SIAM Activity Group on Data Science Career Prize

Oct 23, 2024
UChicago CS News

UChicago CS Researchers Shine at UIST 2024 with Papers, Posters, Workshops and Demonstrations

Oct 10, 2024
UChicago CS News

UChicago Scientists Receive Grant to Expand Global Data Management Platform, Globus

Oct 03, 2024
UChicago CS News

UChicago Researchers Demonstrate the Quantifiable Uniqueness of Former President Donald Trump’s Language Use

Sep 30, 2024
arrow-down-largearrow-left-largearrow-right-large-greyarrow-right-large-yellowarrow-right-largearrow-right-smallbutton-arrowclosedocumentfacebookfacet-arrow-down-whitefacet-arrow-downPage 1CheckedCheckedicon-apple-t5backgroundLayer 1icon-google-t5icon-office365-t5icon-outlook-t5backgroundLayer 1icon-outlookcom-t5backgroundLayer 1icon-yahoo-t5backgroundLayer 1internal-yellowinternalintranetlinkedinlinkoutpauseplaypresentationsearch-bluesearchshareslider-arrow-nextslider-arrow-prevtwittervideoyoutube