SPEAKER 2026
Poonam Goyal
Professor, Birla Institute of Technology & Science
About Talk
Beyond Foundation Models: Building Multimodal AI for the Real World
The rapid progress of large language models and foundation models has transformed artificial intelligence, but many real-world problems cannot be understood through language alone. They require AI systems capable of reasoning across images, text, sensor measurements, temporal signals, and domain knowledge. I will discuss a couple of our recent works on multimodal AI, including models built around LLaVA for vision-language understanding and videoMAE-based architectures for learning from heterogeneous multimodal data from satellite image time series. Through applications spanning visual understanding and domain-specific problems such as agriculture, I will highlight how foundation models can be adapted and extended for real-world settings, where data heterogeneity, efficiency, and trustworthiness become critical. Finally, I will highlight how open-source models and ecosystems are enabling researchers and developers to build increasingly powerful domain-specific AI systems, and discuss opportunities for creating AI that is not only larger, but more grounded, efficient, trustworthy, and useful.
TRACK: Keynote
7th Oct 2026 | Hall A | Time:10:45-11:30
About Speaker
Prof. Poonam Goyal is a Professor in the Department of Computer Science & Information Systems, Birla Institute of Technology & Science, Pilani, Pilani Campus. She is the coordinator of APPCAIR, AI Research Center, BITS Pilani, Pilani Campus. She heads the Web Intelligence and Social Computing Laboratory (WiSOC Lab) of the department. She is also a core member of Advanced Data Analytics and Parallel Technologies Laboratory (ADAPT Lab). Prof. Poonam received her ME degree in Software Systems from BITS Pilani and Ph.D. (Mathematics) from IIT Roorkee.
She specializes in Multimodal learning, Generative AI, Natural Language Processing, Computer Vision and Big Data Analytics. Her research has contributed to various social and scientific domains like Satellite analytics, multi-modal knowledge graphs, Generative AI for design, etc. She has published several research articles in various top-tier conferences and journals such as IEEE Transactions on Multimedia, ACM Transactions on Multimedia Computing Communications and Applications, IEEE Transactions on Social Computing, Expert Systems, AAAI, EACL, IEEE Cluster, IEEE Big Data, IEEE HiPC, IEEE/ACM/ASA DSAA, etc. She has also filed a few patents related to knowledge graphs, OCR, Generative AI for Mechanical Design, Efficient processing of high-resolution SITS, etc. She has served as PI/Co-PI for several sponsored research projects from Governments – Ministry of Education, Ministry of Tribal Affairs, MeitY, VESKI-SMRP, SERB, etc. and from Industry – Adobe, Mercedes-Benz, Micron, etc.
She is a recipient of 2021 Google AI for Social Good research award and 2010 IBM Research Innovation Award under the Smarter Planet Initiative in the area of Scalable Data Analytics. She has been recognized by “India AI” (a Govt. body which is an initiative of MEITY, NeGD, and NASSCOM) as one of the 8 leading women AI researchers in India on International Women’s Day 2021.