OT Bundle of Attachments.zip

This document appears to be a draft pre-solicitation for the "Performance and Reliability Evaluation for Continuous modIfications and uSEability of AI (PRECISE-AI)" program, issued by the National Institutes of Health (NIH) under the Department of Health and Human Services.

The key objectives of the PRECISE-AI program are to create self-correction techniques that can automatically maintain the performance of predictive AI components across diverse clinical settings, and to establish an open-source repository of tools to autonomously maintain the performance of clinical AI Decision Support Tools (AI-DSTs) while enhancing their interpretability and actionability. The program aims to advance novel approaches to support clinician decision-making and scalably manage the performance of AI-DSTs after commercial deployment, including through continuous monitoring, degradation detection, root cause analysis, self-correction, and bidirectional communication with clinicians. The program will test these innovations in real-world settings to demonstrate measurable improvements in clinical decision-making. This pre-solicitation does not describe any specific products or services being procured, pricing terms, award dates, or set-asides. It indicates that a final solicitation will be posted at a later date and that ARPA-H plans to hold a Proposers' Day with additional details.