We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
back
Generating Training Data with Large Language Models w/ Special Guest Marzieh Fadaee
01:16:14
Share
2022/12/13
Neural Search Talks — Zeta Alpha
AI Chapters
Transcribe
Chapters
Introduction
Background and Journey of Marzieh Fadaee
Challenges of Leveraging Large LMs in Information Retrieval
InPars: Motivation and Method
Vanilla vs GBQ Prompting: What's the Difference?
Evaluation and Benchmark: How Does InPars Perform?
Main Results and Takeaways from InPars
Ablations: Prompting, In-Domain vs. MSMARCO Input Documents
Promptagator: Overview and Main Differences with InPars
Retriever Training and Filtering in Promptagator
Main Results from Promptagator
Ablations on Consistency Filtering: Is It the Magic Black-Box Pipeline?
Limitations of Using LMs for Synthetic Data
Future Directions for This Line of Research
Shownotes
Transcript
No transcript made for this episode yet, you may request it for free.