GRPO++: Enhancing Dermatological Reasoning under Low Resource Settings
This work addresses the problem of developing reliable medical AI for dermatology in low-resource settings, though it appears incremental as it builds on existing GRPO and DPO techniques.
The authors tackled the challenge of enabling structured reasoning in vision-language models for dermatology under data scarcity by introducing DermIQ-VLM, which uses a modified GRPO++ method and a multi-stage training pipeline, resulting in notable performance gains over standard fine-tuning approaches on a curated dataset.
Vision-Language Models (VLMs) show promise in medical image analysis, yet their capacity for structured reasoning in complex domains like dermatology is often limited by data scarcity and the high computational cost of advanced training techniques. To address these challenges, we introduce DermIQ-VLM, a VLM developed through a multi-stage, resource-efficient methodology designed to emulate a dermatologist's diagnostic process. Our primary contribution is a modified version of Grouped Relative Policy Optimization (GRPO), called GRPO++, which stabilizes the powerful but data-intensive GRPO framework. Our proposed training pipeline first employs GRPO++ for reasoning-oriented disease recognition, followed by supervised fine-tuning for conversational ability. To mitigate factual errors introduced during this step, we then align the model using Direct Preference Optimization (DPO), leveraging a Knowledge Graph-based system as a scalable proxy for expert preference. A preliminary evaluation on a curated dermatological dataset demonstrates that our proposed methodology yields notable performance gains over standard fine-tuning approaches. These findings validate the potential of our pipeline as a feasible pathway for developing specialized, reliable VLMs in resource-constrained environments.