arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2110.01577cs.LGcs.CY
收藏

从有偏见的毒性标签中学习的实证研究

An Empirical Investigation of Learning from Biased Toxicity Labels

DeepMind(深度思维)

Neel Nanda, Jonathan Uesato, Sven Gowal

摘要:

从人类评分员那里收集注释通常会导致人们希望收集的标签数量和这些标签的质量之间的权衡。因此,通常只能收集少量的高质量标签。在本文中,我们研究了不同的训练策略如何利用人工注释标签的小数据集和合成生成标签的大但嘈杂的数据集(对身份群体表现出偏见)来预测在线评论的毒性。我们评估这些方法的准确性和公平性,以及两者之间的权衡。虽然我们发现对所有数据进行初始训练并对干净数据进行微调会产生具有最高AUC的模型,但我们发现没有一种策略在所有公平性指标上表现最好。

英文摘要:

Collecting annotations from human raters often results in a trade-off between the quantity of labels one wishes to gather and the quality of these labels. As such, it is often only possible to gather a small amount of high-quality labels. In this paper, we study how different training strategies can leverage a small dataset of human-annotated labels and a large but noisy dataset of synthetically generated labels (which exhibit bias against identity groups) for predicting toxicity of online comments. We evaluate the accuracy and fairness properties of these approaches, and trade-offs between the two. While we find that initial training on all of the data and fine-tuning on clean data produces models with the highest AUC, we find that no single strategy performs best across all fairness metrics.

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。

↑