TrustML Event @ UBC, March 2025

Time 

Session

11:30

Registration Opens

12:00-13:00

Welcome, Lunch, and Mingling 

13:00-13:30

Introductions and Poster Briefings

13:30- 14:10

Talk 1 (Virtual, Chair: Julia Rubin)

Eric Martínez, The University of Chicago: "Reasoning in Large Language Models: Insights for and from Human Cognition"

14:10 - 14:40

Tech Briefings (Chair: Mathias Lécuyer)

• Prof. John-Jose Nunez, The University of British Columbia: "A New Method to Interpret Predictive Models in Natural Language Processing Across Multiple Documents"
• Prof. Evan Shelhamer, The University of British Columbia: "Test-Time Adaptation for Robust Visual Recognition"
• Dr. Gargi Mitra, The University of British Columbia: "PRIME: A Comprehensive Threat Modeling Framework for ML-enabled Medical Devices"

14:40 - 15:30

Coffee Break, Posters

• Chun-Yin Huang, The University of British Columbia, "DESA: Overcoming Data and Model Heterogeneities in Decentralized Federated Learning via Synthetic Anchor"
• Wenlong Deng, The University of British Columbia, "DARE the Extreme: Revisiting Delta-Parameter Pruning For Fine-Tuned Models"
• Tina Behnia, The University of British Columbia, "Implicit Geometry of Next-Token Prediction: From Language Sparsity Patterns to Model Representations"
• Masih Beigi Rizi, The University of British Columbia, "Instruction Ambiguity in Large Language Models" (Audience Choice Award)   
• Ritu Thombre, The University of British Columbia, "Quantifying Explainability and Applying it to Quanvolutional Neural Networks"
• Amirreza Soleimanbeigi, The University of British Columbia, "Membership Inference Attack vs Model Inversion Attack in Non-IID Datasets"
• Rudransh Kumar, The University of British Columbia, "Classification using Reinforcement Learning & Semantic-Preserving Transforms (WiP)"
• Sarah Dagger, The University of British Columbia, "Grounding LLM Code Fault Explanations with Symbolic Reasoning"
• Gargi Mitra, The University of British Columbia, "PRIME: A Comprehensive Threat Modeling Framework for ML-enabled Medical Devices"
• Michael Tegegn, The University of British Columbia, "The Effect of Data on ML-based Android Malware Detection"
• Andrew Raitt, The University of Washington, "The Black Box’s Burden: Evaluating Legal Liability Frameworks for Machine Learning Systems"

15:30 - 16:10

Talk 2 (Virtual, Chair: Souradeep Dutta)

Prof. Aziz Z. Huq, The University of Chicago: "The International and Geopolitical Context of AI Regulation"

16:10 - 16:50  

Talk 3 (Virtual, Chair: Simon Oya)

Prof. Reza Shokri, National University of Singapore: "Tracing Data in AI: Auditing Data Privacy with Membership Inference"

16:50 - 17:00

Summary and Closing