Jan 18
CS Tea: Karin de Langis presents "Mind and Machine: Incorporating Cognition into Large Language Models"
Karin de Langis, a PhD candidate at the University of Minnesota will be presenting on her research on Large Language Models (LLMs).
Abstract: State-of-the-art large language models (LLMs) such as Chat-GPT are highly capable of producing fluent, human-like language. In some areas, LLMs produce answers to questions that are comparable to those of human experts. In other areas, LLMs struggle with tasks that young children can accomplish. This discrepancy between LLM and human capabilities is an important area of natural language processing research. This talk discusses research that aims to (1) better understand the gap between LLM and human cognition, and (2) create LLMs that "think" more similarly to humans. Approaches include aligning LLMs to biological signals of human reading, probing LLMs with cognitive assessments originally designed for humans, and embedding LLM systems with specific cognitive strategies.
Bio: Karin is a PhD candidate researching natural language processing at the University of Minnesota Department of Computer Science. She earned her BA in Cognitive Science and Linguistics from Pomona College. Before starting graduate school, she spent three years working as a software engineer.
Join us for cookies, tea and community!
← Return to site Calendar
Go to Campus Calendar →