SEMINAR F’24-EFFICIENT AND ROBUST LANGUAGE ADAPTATION IN MULTILINGUAL LLMS

• Post by: Ashish Agarwal
• October 17, 2024
• Comments off

Presenters: Bibek Upadhata (Ph.D. Candidate SAIL LAB)
Time: Friday 10/11/2024, 4:00 PM – 5:00 pm ET
Place: Maxcy 233A
Recording: EFFICIENT AND ROBUST LANGUAGE ADAPTATION IN MULTILINGUAL LLMS

Bio:

Bibek Upadhayay is a PhD student and Research Assistant at the SAIL Lab, focusing on Multilingualism in Large Language Models, Information Retrieval, NLP, Graph Modeling, and Multi-Modal Deep Learning. Under Dr. Behzadan’s supervision, he has worked on projects extracting actionable intelligence from social media, including a human trafficking detection system currently used by Love Justice International. His thesis tackles fake news classification through style-based and network-centric models, notably improving performance with sentiment-enhanced datasets and a hybrid BERT-Graph Attention Network model, with findings published at IEEE ISI and AAAI. Currently, his research on the security of EEG-based BCIs is featured at IEEE SMC 23.