Please note: This PhD seminar will take place in DC 3301 and online.
Ronak Pradeep, PhD candidate
David R. Cheriton School of Computer Science
Supervisor: Professor Jimmy Lin
Modern information access depends on increasingly complex LLM-driven pipelines, raising challenges in open and effective components, benchmarking, and evaluation. This seminar covers advances in all three fronts. It pushes supervised reranking to its limits, motivating the need for transparent, high-effectiveness components that culminate in RankZephyr, a 7B open listwise reranker matching far larger proprietary models. It then introduces Ragnarök, a standardized end-to-end RAG benchmarking framework and the TREC 2024 RAG Track. Finally, it presents AutoNuggetizer, a scalable nugget-based evaluation methodology for generative systems. Together, these contributions establish an open, reproducible foundation for building and assessing modern information access pipelines.
To attend this PhD seminar in person, please go to DC 3301. You can also attend virtually on Zoom.