BK21 Team have prepared an international seminary for all members and information is the following:
- Title: Integrating Interpretable AI with Emotion and Action Analysis
- Date & time: July 1, 11:00 AM
- Room: MyungShin Bldg. 526 (Meeting room of AI Div.) (명신관 526호 회의실)
- Summary:
In this talk, we will explore two forefront areas of research: Interpretable AI, multimodal
emotion, and human action recognition. The initial segment will be dedicated to the
imperative of comprehending and interpreting machine learning models, particularly focusing
on feature interactions. Here, we'll examine various techniques, such as SHAP and LIME, and
introduce our innovative approach, DnCShap, designed to enhance interpretability in complex
AI systems. The subsequent part will concentrate on multimodal emotion recognition, where
we'll discuss our latest work involving speech, text, and image-based emotion analysis. This
portion will underscore our most recent methodologies and discoveries in identifying and
interpreting emotional cues across diverse modalities, highlighting the importance of
multimodal strategies in understanding human emotions. Additionally, we will delve into the
recent progress in human action recognition. This presentation aims to amalgamate
interpretability with emotion/action recognition, providing insights into how these fields can
synergistically leverage each other's developments
- Lecturer:
Dr. Balasubramanian Raman (Senior Member, IEEE) is Professor (HAG) and Head of the Department of Computer Science & Engineering at the Indian Institute of Technology Roorkee, where he also holds a joint appointment with the Mehta Family School of Data Science & AI. He served as the iHUB Divyasampark Chair Professor from 2022 to 2025. Dr Raman earned his Ph.D. from IIT Madras (2001). His research—spanning machine learning, computer vision, image/video processing, and pattern recognition—has produced more than 250 peer-reviewed publications and a Google-Scholar h-index of 51. He has undertaken post-doctoral and visiting appointments at Rutgers University, the University of Missouri–Columbia, Osaka Metropolitan University, Curtin University, and the University of Cyberjaya. To date, he has supervised 33 Ph.D. scholars (with a further 16 in progress) and regularly directs large, multi-institutional programmes funded by leading national agencies and industry partners. An inventor of an Indian patent for a real-time fog-removal imaging system, he has also co-authored two books—Deep Learning Through the Prism of Tensors (Springer Nature, 2024) and The Geometry of Intelligence: Foundations of Transformer Networks in Deep Learning (Springer Nature, 2025). His honours include the DST BOYSCAST Fellowship, two IIT Roorkee Outstanding Teacher Awards, the Ramkumar Prize for Outstanding Teaching and Research, and the ICPC Coach Award, which recognised his teams’ top-50 finishes at the ACM ICPC World Finals.
