Embedded Files
9:00-9:10 AM
Opening Remarks
9:10-10:00 AM
Tutorial: Jaden Fiotto-Kaufman (NDIF)
NDIF tutorial: Toward a Shared Stack for Vision Interpretability
10:00-10:30 AM
Coffee Break
11:00-11:30 AM
Young Researcher Talk: Dana Arad
Through Layers and Tokens: Tracing Information Flow in Vision and Language Models
11:30 AM-11:50 PM
Spotlight Talks (2 x 10 min)
A Case Study on Hidden Bias in Vision-Language Model Activations
Feature Alignment for Scalable B-cosification of Foundational Vision Transformers
11:50-2:30 PM
Lunch Break
12:00-2:30 PM
Poster Session (326 - 337 exhibit hall A)
3:00-3:30 AM
Invited Talk: Tsui-Wei (Lily) Weng
Toward a Science of Interpretability: Structured Explanations and Reliable Evaluation
3:30-4:00 PM
Coffee Break
4:00-4:30 PM
Invited Talk: Zhuang Liu
How Do Vision Models Learn to Reason? - Lessons from Open Visual RL
4:30-4:50 PM
Spotlight Talks (2 x 10 min)
FaCT: Faithful Concept Traces for Explaining Neural Networks
LogitDynamics: Reliable ViT Error Detection from Layerwise Logit Trajectories
4:50-5:00 PM
Closing Remarks
Page updated
Google Sites
Report abuse