log in  |  register  |  feedback?  |  help  |  web accessibility
Robust, Fair, and Culturally-Aware Commonsense Reasoning for NLP
Rachel Rudinger
IRB 4105 or https://umd.zoom.us/j/93754397716?pwd=GuzthRJybpRS8HOidKRoXWcFV7sC4c.1
Wednesday, October 1, 2025, 3:00-4:00 pm
  • You are subscribed to this talk through .
  • You are watching this talk through .
  • You are subscribed to this talk. (unsubscribe, watch)
  • You are watching this talk. (unwatch, subscribe)
  • You are not subscribed to this talk. (watch, subscribe)
Abstract

Successful communication in natural language requires common ground: that is, a mutual understanding of conversation partners’ beliefs, intents, and conversational goals. For large language models (LLMs) to build useful mental models of their human users, they must possess certain human-like cognitive abilities, including basic knowledge of the world (i.e., “common sense”) and the ability to draw inferences about people or situations from partial observations (i.e., reasoning under uncertainty). However, valid measurement of these abilities in LLMs is not a simple task. In the first part of this talk, I will discuss some of my lab’s recent work developing methods to analyze the robustness of these reasoning capabilities in LLMs beyond traditional benchmark accuracy. Drawing on theories from social psychology and cultural anthropology, the second part of the talk will present work demonstrating socio-cultural biases in how LLMs reason about people, raising fairness and safety concerns about their use in certain contexts.

Bio

Rachel Rudinger is an Assistant Professor in Computer Science, with appointments in the Institute for Advanced Computer Science (UMIACS) and Linguistics. She is a member of the Computational Linguistics and Information Processing (CLIP) Lab at UMD. Her research interests lie in the areas of natural language understanding, commonsense reasoning, computational semantics/pragmatics, and issues of sociocultural fairness in NLP systems. Rachel has published over 60 papers in top Natural Language Processing conferences and journals, and is a 2024 recipient of the National Science Foundation CAREER Award. She holds a Ph.D. in Computer Science from Johns Hopkins University.

This talk is organized by Samuel Malede Zewdu