Ran Gu

Forget static predictions: OcularChat brings clinically meaningful dialogue and reasoning to AI-powered AMD diagnosis, outperforming existing models and impressing ophthalmologists.

Chain-of-Thought prompting doesn't always improve LLMs' ability to solve discrete optimization problems, and surprisingly, "disordered" datasets can sometimes boost performance on simpler tasks.

CT-Bench reveals that even state-of-the-art multimodal models struggle with lesion understanding in CT scans, highlighting the need for specialized datasets and fine-tuning to bridge the gap between AI and radiologist performance.