Towards More Explainable Artificial Intelligence for Electrocardiogram Analysis
As part of the CVN–MICS seminar series, CentraleSupélec will welcome Srinivasan Parthasarathy, Professor at The Ohio State University (USA), on Wednesday, November 4, 2026. He is currently a visiting professor with the support of the DataIA Paris-Saclay Institute.
His talk, entitled “Explainable, Efficient, Effective, and Auditable ECG-Based Cardiac Condition Classification through Visibility Graph Neural Networks”, will explore innovative AI approaches to improving the classification of cardiac conditions from electrocardiograms (ECGs), while enhancing the transparency and interpretability of the results.
Deep learning models applied to ECG analysis now achieve remarkable performance in detecting cardiac conditions. However, identifying the specific features of the signals that influence their predictions often remains challenging. To address this issue, Srinivasan Parthasarathy will present LYNX-ECG, an approach based on Visibility Graph Neural Networks. This method aims to combine classification performance, computational efficiency, and explainability, preserving essential ECG signal information while significantly reducing storage requirements.
This research highlights the potential of innovative AI approaches to develop more transparent and interpretable diagnostic support tools that are better aligned with clinical practice.
About the Speaker
Srinivasan Parthasarathy is a Professor and University Distinguished Scholar at The Ohio State University. His research focuses on data analysis, machine learning, graph analytics, and responsible artificial intelligence.
Recognized for his major contributions to these fields, he has received numerous scientific awards and distinctions. He is notably an IEEE Fellow and Co-Director of the Center for Responsible AI and Governance.
His research has applications across various sectors, particularly healthcare, and contributes to the development of more reliable and responsible AI systems.
Practical Information
- Date: Wednesday, November 4, 2026
- Time: 10:00 AM
- Venue: Amphitheater V – Eiffel Building, CentraleSupélec
- Address: 3 rue Joliot-Curie, 91192 Gif-sur-Yvette, France
- Speaker: Srinivasan Parthasarathy, The Ohio State University
- Organized as part of: CVN–MICS Seminar Series
- Supported by: DataIA Paris-Saclay Institute, through its Visiting Professor Program