Continually Improving Extractive QA via Human Feedback
This work addresses the challenge of continuously enhancing QA systems for information-seeking users, though it appears incremental as it builds on existing feedback-based learning methods.
The researchers tackled the problem of improving extractive question answering systems over time by incorporating human user feedback, demonstrating that their iterative approach leads to effective model improvement across different data regimes and shows significant potential for domain adaptation.
We study continually improving an extractive question answering (QA) system via human user feedback. We design and deploy an iterative approach, where information-seeking users ask questions, receive model-predicted answers, and provide feedback. We conduct experiments involving thousands of user interactions under diverse setups to broaden the understanding of learning from feedback over time. Our experiments show effective improvement from user feedback of extractive QA models over time across different data regimes, including significant potential for domain adaptation.