User queries in natural settings, such as “provide a design for a disk topology for a NAS built on TrueNAS Scale, as well as a dataset layout,” differ significantly from those produced in crowdsourced environments, like “Did Aristotle use a laptop?” Addressing real-world user queries requires language models (LMs) to possess more comprehensive abilities, such as resolving ambiguity, applying world knowledge, and reasoning within context. This talk will cover evaluating and training LMs to reason in real-world scenarios.
Speaker Bio
Wenting Zhao is a final-year Ph.D. candidate in Computer Science at Cornell University, advised by Professors Claire Cardie and Professor Sasha Rush. Her research focuses on reasoning: she develops techniques that can accurately reason over real-world scenarios and creates benchmarks that reliably reflect the reasoning performance of language models when deployed in the real world. She was named the intern of the year at AI2, and she organized reasoning tutorials and workshops at ACL conferences.
More Details
- When: Wed 20 Nov 2024, at 1 - 2 pm (Brisbane time)
- Speaker: Wenting Zhao (Cornell University)
- Host: Ruihong Qiu
- Venue: Online
- Zoom: https://uqz.zoom.us/j/82860997780
No.24-14 Long-Range Meets Scalability: Unveiling a Linear-Time Graph Neural Network for Recommendation at Scale