PathCoT: Chain-of-Thought Prompting for Zero-shot Pathology Visual Reasoning
This addresses a domain-specific problem for pathology AI applications, with incremental improvements in method adaptation.
The paper tackles the problem of multimodal large language models (MLLMs) underperforming on pathology visual reasoning tasks due to lack of domain-specific knowledge and errors in chain-of-thought (CoT) reasoning, proposing PathCoT, a zero-shot CoT prompting method that integrates expert knowledge and self-evaluation, which shows effectiveness on the PathMMU dataset.
With the development of generative artificial intelligence and instruction tuning techniques, multimodal large language models (MLLMs) have made impressive progress on general reasoning tasks. Benefiting from the chain-of-thought (CoT) methodology, MLLMs can solve the visual reasoning problem step-by-step. However, existing MLLMs still face significant challenges when applied to pathology visual reasoning tasks: (1) LLMs often underperforms because they lack domain-specific information, which can lead to model hallucinations. (2) The additional reasoning steps in CoT may introduce errors, leading to the divergence of answers. To address these limitations, we propose PathCoT, a novel zero-shot CoT prompting method which integrates the pathology expert-knowledge into the reasoning process of MLLMs and incorporates self-evaluation to mitigate divergence of answers. Specifically, PathCoT guides the MLLM with prior knowledge to perform as pathology experts, and provides comprehensive analysis of the image with their domain-specific knowledge. By incorporating the experts' knowledge, PathCoT can obtain the answers with CoT reasoning. Furthermore, PathCoT incorporates a self-evaluation step that assesses both the results generated directly by MLLMs and those derived through CoT, finally determining the reliable answer. The experimental results on the PathMMU dataset demonstrate the effectiveness of our method on pathology visual understanding and reasoning.