About Us
The Vidal Lab develops mathematical foundations for machine learning, computer vision, and dynamical systems. Our research spans deep learning theory, trustworthy and parsimonious representation learning, optimization, and continual learning. In computer vision, we work on 3D vision, video and image understanding, and vision-language models. We also apply these methods to biomedical data science, including medical image analysis and computer vision for health. Additionally, we study the theory of dynamical systems, including hybrid, multi-agent, and linear systems.
Recent News
- Feb 2026 New preprints on Hierarchical Concept Pursuit and ImageRAGTurbo.
- Jan 2025 Paper on Gradient Flow accepted to ICML 2025.
- Nov 2024 Four papers accepted to NeurIPS 2024!
- Jul 2024 Paper on active binary testing accepted to ICML 2024.
Project Highlights
MICCAI 2025
IP-CRR: Information Pursuit for Interpretable Classification of Chest Radiology Reports
Preprint 2025
Conformal Information Pursuit for Interactively Guiding Large Language Models
Preprint 2025
SECA: Semantically Equivalent and Coherent Attacks for Eliciting LLM Hallucinations
IEEE TBE 2025