InclusionMed, the world's first independent evaluation and analysis platform for healthcare AI, has officially launched. The launch includes the simultaneous release of the InclusionMed Atlas, a healthcare AI evaluation map, and the inaugural InclusionMed Index, a healthcare AI evaluation index, designed to provide a systematic assessment framework for healthcare AI. The platform is jointly developed by researchers from 29 institutions in China and abroad, addressing a global gap in comprehensive evaluation of healthcare AI. To tackle the limitation that existing single-task evaluation benchmarks cannot fully capture model performance across multiple scenarios, the platform employs a nine-grid evaluation map to outline evaluation coverage, while the index aggregates evaluation results to present models' overall performance and their relative strengths and weaknesses across tasks. The inaugural evaluation covers more than a thousand tasks across two major domains, which were screened to form the first index. The project team will subsequently collaborate with various stakeholders to develop additional evaluation tasks, further examining AI performance across multiple types of healthcare scenarios.
