Vision-Language Models

CURE: Curriculum-Guided Multi-Task Training for Reliable Anatomy-Grounded Report Generation

We propose CURE, a curriculum-guided multi-task training framework for reliable, anatomy-grounded radiology report generation. Developed with collaborators at KAUST and Pontificia Universidad Católica de Chile, CURE was selected for oral presentation …

SAVeS: Steering Safety Judgments in Vision-Language Models via Semantic Cues

We study how semantic cues can be used to steer safety judgments in vision-language models, proposing SAVeS as a method to improve the reliability of safety-relevant decisions made by multimodal models.