IIT Kanpur, Kanpur
Nisheeth Srivastava is a Professor of Common Sense in the Department of Computer Science & Engineering and the Department of Cognitive Science at the Indian Institute of Technology Kanpur, Kanpur. He completed his postdoctoral training at UC San Diego and received his PhD from the University of Minnesota, USA. He currently leads CDIS (Center for Developing Intelligent Systems) and ARF (Airawat Research Foundation), conducting interdisciplinary research at the intersection of artificial intelligence, cognitive science, and computational social systems.
Session 2A: Symposium on Artificial Intelligence in Health
Chairperson: Somdatta Sinha, IISER Kolkata, Nadia and Narendra Dixit, IISc, Bengaluru
BODH – A New Way of Validating AI Models for Health
BODH AI has been conceived by our team at IIT Kanpur as a platform for benchmarking, testing, and validating AI algorithms for clinical use in a privacy-preserving and institutionally credible manner. The framework recognizes that conventional software testing is inadequate for healthcare AI, that robust third-party validation is difficult because of privacy constraints and limited access to diverse clinical datasets, and that public institutions require dedicated operational capacity to manage such testing systems. The platform architecture is built around an open benchmarking mechanism, consent-governed data use, and quality-preserving management of benchmark datasets. The same framework further envisages a structured ecosystem in which model developers can test algorithms in a double-blind, federated setup, benchmark performance on out-of-sample data, and generate an auditable record of model behavior, while preserving privacy and statistical validity. This talk will introduce the audience to BODH, and describe how it works in a lightly technical manner.