Effective Transfer Learning for Identifying Similar Questions - Matching User Questions to COVID-19 FAQs

Explore effective transfer learning techniques for identifying similar questions in this 20-minute conference talk from KDD 2020. Delve into the challenges and approaches for matching user questions to COVID-19 FAQs, presented by Clara McCreery, Namit Katariya, Anitha Kannan, Manish Chablani, and Xavier Amatriain. Learn about domain-similarity vs task similarity, error analysis, and understanding model errors. Discover how these techniques are applied in practice to match user COVID-questions to FAQs and see the product in action. Gain insights into the contributions made in this field of automated FAQ matching, particularly in the context of the COVID-19 pandemic.

Intro
Motivation
COVID-19: Automated FAQ Matching
Challenges
Our approach
Underlying Models
Approach Overview
Domain-similarity vs Task similarity
Error Analysis
Understanding Model Errors
Matching User COVID-questions to FAQS
Product in action
Contributions