<ul><li>Introduction of QoQ-Med-7B/32B, the first open generalist clinical foundation model that reasons across medical images, time-series signals, and text reports.</li><li>Trained with Domain-aware Relative Policy Optimization (DRPO) to address performance imbalance from skewed clinical data distributions.</li><li>Demonstrated improved diagnostic performance by 43% in macro-F1 on average across all visual domains compared to other training methods like GRPO.</li><li>QoQ-Med highlights salient regions related to diagnosis with an IoU 10x higher than open models and reaches the performance of OpenAI o4-mini, aiming to foster reproducibility in research.</li></ul>

QoQ-Med: Building Multimodal Clinical Foundation Models with Domain-Aware GRPO Training

Discover more