Talk title:
The Art of (Artificial) Reasoning
Event details:
Date: Thursday, April 16, 2025
Time: 5-7 p.m.
Location: Schwartz Reisman Innovation Campus, Room TBC
108 College Street, Toronto, ON M5G 0C6
This lecture is open to the public. Registration details coming soon.
Abstract:
Scaling laws suggest that “more is more” — brute-force scaling of data and compute leads to stronger AI capabilities. However, despite rapid progress on benchmarks, state-of-the-art models still exhibit "jagged intelligence," indicating that current scaling approaches may have limitations in terms of sustainability and robustness. Additionally, while the volume of papers on arXiv continues to grow rapidly, our scientific understanding of artificial intelligence hasn't kept pace with engineering advances, and the current literature presents seemingly contradictory findings that can be difficult to reconcile. In this talk, I will discuss key insights into the strengths and limitations of LLMs, examine when reinforcement learning succeeds or struggles in reasoning tasks, and explore methods for enhancing reasoning capabilities in smaller language models to help them close the gap against their larger counterparts in specific domains.
Bio:
Yejin Choi is the Dieter Schwarz Foundation Professor and Senior Fellow at the Department of Computer Science at Stanford University and the Stanford Institute for Human-Centered Artificial Intelligence (HAI) respectively. She is currently Senior Director at NVIDIA, and was previously Professor at UW and Senior Director at AI2. Choi is a MacArthur Fellow (class of 2022), AI2050 Senior Fellow (class of 2024), and named among Time100's Most Influential People in AI in 2023 and 2025. In addition, Choi is a co-recipient of 2 Test-of-Time awards and 10 Best and Outstanding Paper Awards at top AI conferences including ACL, ICML, NeurIPS, ICCV, CVPR, and AAAI, the Borg Early Career Award (BECA) in 2018, the inaugural Alexa Prize Challenge in 2017, and IEEE AI’s 10 to Watch in 2016. Choi was a main stage speaker at TED 2023, and a keynote speaker for a dozen conferences across several AI disciplines including ACL, CVPR, ICLR, MLSys, VLDB, WebConf, and AAAI. Her current research interests include fundamental limits and capabilities of large language models, alternative training recipes for language models, symbolic methods for neural networks, reasoning and knowledge discovery, moral norms and values, pluralistic alignment, and AI safety.