Bio
Srijan Kumar is an Assistant Professor at the College of Computing at the Georgia Institute of Technology. He completed his postdoctoral training at Stanford University, received a Ph.D. and M.S. in Computer Science from the University of Maryland, College Park, and B.Tech. from the Indian Institute of Technology, Kharagpur. He develops Data Mining methods to detect and mitigate the pressing threats posed by malicious actors (e.g., evaders, sockpuppets, etc.) and harmful content (e.g., misinformation, hate speech etc.) to web users and platforms. His methods have been used in production at Flipkart (India’s largest e-commerce platform) and influenced Twitter’s Birdwatch system. He has been selected as an NSF CAREER awardee, a Kavli Fellow by the National Academy of Sciences, named as Forbes 30 under 30 honoree in Science, ACM SIGKDD Doctoral Dissertation Award runner-up 2018, and best paper honorable mention award from the ACM Web Conference. He has also won the Adobe Faculty Award and the Facebook Faculty Award. His research has been covered in the popular press, including CNN, The Wall Street Journal, Wired, and New York Magazine.
Talk Title : Robustness of LLMs and Multimodal Models
Abstract: LLMs and multimodal models make unexpected errors. I will discuss my group’s recent research (ACL 2023, EMNLP 2022, ICWSM 2022) that investigates the scenarios in which these errors occur in multilingual, multimodal, and few-shot learning settings.