Continually Improving Extractive QA via Human Feedback

Ge Gao, Hung-Ting Chen, Yoav Artzi, Eunsol Choi
5/21/2023
cs.CLcs.AIcs.LG

No official code found. Best implementation: lil-lab/qa-from-hf (9 stars, last commit 2023-10-10).

Abstract

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.

View on arXivView PDF

Code Implementations(1)

lil-lab/qa-from-hfCommunity100%
92Shell, PythonApr 13, 20233 years ago

Similar Papers

IQA-EVAL: Automatic Evaluation of Human-Model Interactive Question Answering

83% match2026

Learning from Natural Language Feedback for Personalized Question Answering

81% match2025

Reinforcement learning for question answering in programming domain using public community scoring as a human feedback

80% match2024

Aligning LLMs through Multi-perspective User Preference Ranking-based Feedback for Programming Question Answering

80% match2024

AirQA: A Comprehensive QA Dataset for AI Research with Instance-Level Evaluation

80% match1 repo72025

Learning Answer Generation using Supervision from Automatic Question Answering Evaluators

79% match2023

Cite this paper

@article{gao2023continually,
  title  = {Continually Improving Extractive QA via Human Feedback},
  author = {Ge Gao and Hung-Ting Chen and Yoav Artzi and Eunsol Choi},
  year   = {2023},
  eprint = {2305.12473},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.12473v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.12473.svg)](https://codeofpaper.com/paper/2305.12473)

Discussion