Tuesday 13 January 2026 Host: Ajitha RajanSpeakers: Ajitha Rajan and Piyush BoroleTitle: Designing Transparent and Clinically Interpretable AI for Chest X-Rays and ImmunoinformaticsAbstract: In recent years, deep learning has driven major advances in medicine from diagnosing patient scans, predicting protein functions and designing new proteins to modeling immune responses and uncovering complex patterns in DNA and RNA with foundation models. Despite their remarkable accuracy, these models often lack interpretability—a key requirement for clinical reliability and adoption. This talk will examine these challenges and outline potential solutions, with a focus on clinical interpretability for chest x-ray diagnosis and predicting immune responses through antigen presentation on MHC Class I molecules. Ajitha will present the Chest radiology contribution and Piyush will discuss the development of a transparent and interpretable model for MHC-I antigen presentation.Biography: Professor Ajitha Rajan is Chair of Software Testing and Verification in the School of Informatics at the University of Edinburgh. In addition to software testing and verification, her research also explores clinically interpretable and robust AI techniques for cancer diagnosis and treatment. Dr. Piyush Borole earned his PhD from the ANC (IML) Institute in September 2025, specializing in explainable biomedical AI. He is currently a Data Scientist at the Data Intelligence Hub, School of Informatics. His interests include interpretable models and trustworthy AI systems. Jan 13 2026 13.00 - 14.00 Tuesday 13 January 2026 Speakers: Ajitha Rajan and Piyush Borole IF, G.03
Tuesday 13 January 2026 Host: Ajitha RajanSpeakers: Ajitha Rajan and Piyush BoroleTitle: Designing Transparent and Clinically Interpretable AI for Chest X-Rays and ImmunoinformaticsAbstract: In recent years, deep learning has driven major advances in medicine from diagnosing patient scans, predicting protein functions and designing new proteins to modeling immune responses and uncovering complex patterns in DNA and RNA with foundation models. Despite their remarkable accuracy, these models often lack interpretability—a key requirement for clinical reliability and adoption. This talk will examine these challenges and outline potential solutions, with a focus on clinical interpretability for chest x-ray diagnosis and predicting immune responses through antigen presentation on MHC Class I molecules. Ajitha will present the Chest radiology contribution and Piyush will discuss the development of a transparent and interpretable model for MHC-I antigen presentation.Biography: Professor Ajitha Rajan is Chair of Software Testing and Verification in the School of Informatics at the University of Edinburgh. In addition to software testing and verification, her research also explores clinically interpretable and robust AI techniques for cancer diagnosis and treatment. Dr. Piyush Borole earned his PhD from the ANC (IML) Institute in September 2025, specializing in explainable biomedical AI. He is currently a Data Scientist at the Data Intelligence Hub, School of Informatics. His interests include interpretable models and trustworthy AI systems. Jan 13 2026 13.00 - 14.00 Tuesday 13 January 2026 Speakers: Ajitha Rajan and Piyush Borole IF, G.03
