arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.12005cs.HCcs.LGstat.ML
收藏

再探机器学习信任可视化:2023年领域现状

Visualization for Trust in Machine Learning Revisited: The State of the Field in 2023

Northwestern University(西北大学);Linköping University(林雪平大学);Linnaeus University(林奈大学)

Angelos Chatzimparmpas, Kostiantyn Kucher, Andreas Kerren

AI总结:

该研究更新了机器学习信任可视化领域的现状调研,将收录技术从200项扩充至542项,分析2023年秋季前的领域趋势与8项开放挑战,证实该方向近年快速增长,可视化可助力提升模型可解释性与检验新架构。

摘要:

用于可解释、可信赖机器学习的可视化仍是信息可视化与视觉分析领域最重要、研究最广泛的方向之一,应用于医学、金融、生物信息学等多个领域。继我们2020年涵盖200项技术的现状报告后,我们持续收集描述可视化技术的同行评议论文,沿用此前建立的含119个类别的分类体系对其分类,并将收集到的542项技术在在线调研浏览器中开放。在这篇综述中,我们呈现截至2023年秋季对该数据集开展新分析的更新发现,探讨机器学习中可视化应用的趋势、洞见与8项开放挑战。研究结果证实,过去三年间,用于提升机器学习模型信任度的可视化技术呈快速增长趋势,例如可视化可助力改进主流模型可解释性方法、检验新型深度学习架构。

英文摘要:

Visualization for explainable and trustworthy machine learning remains one of the most important and heavily researched fields within information visualization and visual analytics with various application domains, such as medicine, finance, and bioinformatics. After our 2020 state-of-the-art report comprising 200 techniques, we have persistently collected peer-reviewed articles describing visualization techniques, categorized them based on the previously established categorization schema consisting of 119 categories, and provided the resulting collection of 542 techniques in an online survey browser. In this survey article, we present the updated findings of new analyses of this dataset as of fall 2023 and discuss trends, insights, and eight open challenges for using visualizations in machine learning. Our results corroborate the rapidly growing trend of visualization techniques for increasing trust in machine learning models in the past three years, with visualization found to help improve popular model explainability methods and check new deep learning architectures, for instance.

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。

↑