arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2109.08131cs.HCcs.CYcs.LG
收藏

研究机器学习数据:当我们意味着力量时,为什么要谈论偏见?

Studying Up Machine Learning Data: Why Talk About Bias When We Mean Power?

Technische Universität Berlin(柏林工业大学);Weizenbaum Institute(魏岑鲍姆研究所);University of Toronto(多伦多大学);Schwartz Reisman Institute(施瓦茨·赖斯曼研究所)

Milagros Miceli, Julian Posada, Tianling Yang

摘要:

机器学习(ML)的研究主要认为,在不完整或有偏见的数据集上训练的模型可能会导致歧视性的输出。在这篇评论中,我们建议通过采用功率感知的视角来“研究”ML数据集,将研究重点从面向偏见的框架中转移出来。这意味着要考虑到历史上的不平等、劳动条件和数据中的认识论立场。我们利用HCI和CSCW的工作来支持我们的论点,批判性地分析以前的研究,并指出在我们的社区内的两个共存的工作线-一个偏向,其他的权力意识。因此,我们强调需要在三个领域进行对话与合作:数据质量、数据工作和数据记录。在第一个领域,我们认为,减少社会问题的“偏见”错过了基于上下文的数据的性质。在第二部分中,我们强调了数据工作者的劳动所涉及的企业力量和市场需求,这些数据工作者随后塑造了ML数据集。最后,我们建议扩大目前在数据集文档中以数据一致性为导向的努力,以反映数据设计和生产的社会背景。

英文摘要:

Research in machine learning (ML) has primarily argued that models trained on incomplete or biased datasets can lead to discriminatory outputs. In this commentary, we propose moving the research focus beyond bias-oriented framings by adopting a power-aware perspective to "study up" ML datasets. This means accounting for historical inequities, labor conditions, and epistemological standpoints inscribed in data. We draw on HCI and CSCW work to support our argument, critically analyze previous research, and point at two co-existing lines of work within our community -- one bias-oriented, the other power-aware. This way, we highlight the need for dialogue and cooperation in three areas: data quality, data work, and data documentation. In the first area, we argue that reducing societal problems to "bias" misses the context-based nature of data. In the second one, we highlight the corporate forces and market imperatives involved in the labor of data workers that subsequently shape ML datasets. Finally, we propose expanding current transparency-oriented efforts in dataset documentation to reflect the social contexts of data design and production.

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。

↑