.png){kind=icon}
返回
PredDiff: Explanations and interactions from conditional expectations
.png){kind=icon}
OA
AI
DOI：10.1016/j.artint.2022.103774.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
PredDiff is a model-agnostic, local attribution method that is firmly rooted in probability theory. Its simple intuition is to measure prediction changes while marginalizing features. In this work, we clarify properties of PredDiff and its close connection to Shapley values. We stress important differences between classification and regression, which require a specific treatment within both formalisms. We extend PredDiff by introducing a new, well-founded measure for interaction effects between arbitrary feature subsets. The study of interaction effects represents an inevitable step towards a comprehensive understanding of black-box models and is particularly important for science applications. Equipped with our novel interaction measure, PredDiff is a promising model-agnostic approach for obtaining reliable, numerically inexpensive and theoretically sound attributions. (c) 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Keyword:
Explainable AI
Interactions
Feature attribution
Interpretability
Shapley values
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
13.9
论文数:
6.1K
被引数:
1.9W
机构
引用论文
Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications解释深度神经网络及其以后: 方法和应用综述.png)
PROCEEDINGS OF THE IEEEOAAI
PROCEEDINGS OF THE IEEEIF25.9
OAAI分享收藏Explaining individual predictions when features are dependent: More accurate approximations to Shapley values当特征依赖时解释单个预测: 更准确地逼近Shapley值OAAI
OAAI分享收藏The response of rectangular piezoelectric sensors to Rayleigh and Lamb ultrasonic waves矩形压电传感器对瑞利和兰姆超声波的响应OAAI
OAAI分享收藏Unmasking Clever Hans predictors and assessing what machines really learn揭开聪明的汉斯预测器并评估机器真正学到的东西NATURE COMMUNICATIONSOAAI
NATURE COMMUNICATIONSIF15.7
OAAI分享收藏Health Literacy – a review of research using the European Health Literacy Questionnaire (HLS-EU-Q16) in 2010-2018健康素养-2010-2018使用欧洲健康素养问卷 (HLS-EU-Q16) 进行的研究综述
OAAI
OAAI分享收藏Recyclable graphite oxide catalyzed Friedel–Crafts addition of indoles to α,β-unsaturated ketones
PREAI
分享收藏
查看更多内容