arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2108.01764cs.CLcs.AI
收藏

Q-Pain:一个测量疼痛管理中社会偏见的问答数据集

Q-Pain: A Question Answering Dataset to Measure Social Bias in Pain Management

Stanford University(斯坦福大学);Harvard University(哈佛大学)

Cécile Logé, Emily Ross, David Yaw Amoah Dadey, Saahil Jain, Adriel Saporta, Andrew Y. Ng, Pranav Rajpurkar

摘要:

自然语言处理(NLP)的最新进展,特别是自动化问答(QA)系统,已经证明了令人印象深刻的语言流畅性和反映社会偏见的有害趋势。在这项研究中,我们介绍了Q-Pain,这是一个用于评估疼痛管理背景下医疗QA偏差的数据集,疼痛管理是临床决策中最具挑战性的形式之一。随着数据集,我们提出了一个新的,严格的框架,包括一个样本实验设计,以衡量潜在的偏见时,作出治疗决定。我们通过评估两个参考人工智能系统GPT-2和GPT-3来展示其使用,并发现交叉种族-性别亚组之间的治疗存在统计学显著差异,从而重申了人工智能在医疗环境中带来的风险,以及在部署医疗人工智能应用程序之前需要像我们这样的数据集来确保安全性。

英文摘要:

Recent advances in Natural Language Processing (NLP), and specifically automated Question Answering (QA) systems, have demonstrated both impressive linguistic fluency and a pernicious tendency to reflect social biases. In this study, we introduce Q-Pain, a dataset for assessing bias in medical QA in the context of pain management, one of the most challenging forms of clinical decision-making. Along with the dataset, we propose a new, rigorous framework, including a sample experimental design, to measure the potential biases present when making treatment decisions. We demonstrate its use by assessing two reference Question-Answering systems, GPT-2 and GPT-3, and find statistically significant differences in treatment between intersectional race-gender subgroups, thus reaffirming the risks posed by AI in medical settings, and the need for datasets like ours to ensure safety before medical AI applications are deployed.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑