Workshop Program

Time Session
08:50 - 09:00 Welcome
09:00 - 09:40 Invited Talk by Aleksandr Petiushko
Safety and Uncertainty in Autonomous Driving
09:40 - 10:20 Invited Talk by Sharon Li
A Representative-Centric Perspective for Hallucination Detection
10:20 - 10:30 Coffee Break
10:30 - 11:10 Invited Talk by Bolei Zhou
Learning Safety-aware Agents from Human Feedback and Adversary Environments
11:10 - 11:50 Invited Talk by Arash Vahdat
Rare Yet Real: Generative Modeling Beyond the Modes
11:50 - 13:00 Lunch
13:00 - 13:30 Contributed Talks (Full Papers)
  • Improving Optical Flow and Stereo Depth Estimation by understanding learning difficulties
  • D-Feat Occlusions: Diffusion Features for Robustness to Partial Visual Occlusions in Object Recognition
  • The Surprising Utility of Group Partitioning in Improving Conformal Prediction of Visual Classifiers under Distributional Shifts
13:30 - 14:10 Invited Talk by Judy Hoffman
Bridging the Gap: Robust Robot Learning from Human Video
14:10 - 14:30 Contributed Spotlights
  • View-Dependent Uncertainty Estimation of 3D Gaussian Splatting
  • Spatially-Aware Evaluation of Segmentation Uncertainty
  • Compositional Targeted Multi-Label Universal Perturbations
  • Quantile UQNet: A Conformal, Nonlinear Scaling Model for Uncertainty Prediction
  • Centaur: Robust End-to-End Autonomous Driving with Test-Time Training
  • Consensus-Driven Active Model Selection
  • Uncertainty Quantification for Gradient-based Explanations in Neural Networks
  • WQLCP: Weighted Adaptive Conformal Prediction for Robust Uncertainty Quantification Under Distribution Shifts
  • HARMONY: Hidden Activation Representations and Model Output-Aware Uncertainty Estimation for Vision-Language Models
14:30 - 15:50 Poster Session & Coffee Break
15:50 - 16:30 Invited Talk by Jindong Gu
Toward the Risks Brought by Visual Input into Multimodal LLMs
16:30 - 17:10 Invited Talk by Michael Kirchhof
(De/Re)constructing uncertainties for generative vision and language
17:10 - 17:20 Closing Remarks