vered talk

Author: logan-siyao-peng

_events/2025-12-16-vered-shwartz.md

@@ -0,0 +1,14 @@
+---
+title: "Towards Slightly Less Bad Models: Reliability, Fairness, and Generalization in LLMs and VLMs"
+abstract: "<p>Despite remarkable progress and widespread adoption of generative AI tools such as LLMs and VLMs, fundamental challenges remain for ensuring reliability, fairness, and generalization for real-world usage. In this talk, I will discuss research on systematically evaluating and improving the capabilities of models along these dimensions. With respect to reliability, I will present our ongoing work on detecting and resolving uncertainty in legal and medical applications. For fairness, I will discuss LLMs’ cultural bias and how we can make them more cultural competent. Finally, I will present two lines of work testing models generalizability to new contexts — identifying surprising events in videos models and crosslingual knowledge transfer in LLMs.</p>"
+speaker: Vered Shwartz <br/>(Assistant Professor of Computer Science at the University of British Columbia)
+bio: '<p>Vered is an Assistant Professor of Computer Science at the University of British Columbia, and a CIFAR AI Chair at the Vector Institute. Her research interests focus on natural language processing, with the fundamental goal of building models capable of human-level understanding of natural language. Her lab is currently working on testing and improving the capabilities of large language models and vision and language models, developing culturally-competent AI, and responsible NLP applications in sensitive domains (e.g., legal, medical).</p><p>Before joining UBC, Vered was a postdoctoral researcher at the Allen Institute for AI (AI2) and the Paul G. Allen School of Computer Science & Engineering at the University of Washington. Prior to that, she did her PhD (2019), M.Sc. (2015), and B.Sc. (2013) in Computer Science in Bar-Ilan University. She was part of the Natural Language Processing lab where she worked on lexical and compositional semantics.</p><p>Vered is a member of the UBC NLP Group, CAIDA, Language Sciences, and ICICS.</p>'
+website: https://www.cs.ubc.ca/~vshwartz/
+time: December 16, 2025; 10:00-11:00
+location: Akademiestr. 7, room 218A (meeting room)
+roomfinder: https://mainlp.github.io/contact/
+img: assets/img/vered-shwartz.jpg
+imgalt: Portrait of Vered Shwartz
+imgside: right
+anchor: 2025-12-16-vered-shwartz
+---