arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2011.04328cs.LGcs.AI
收藏

机器学习模型的风险评估

Risk Assessment for Machine Learning Models

Neurocat GmbH(Neurocat有限公司);Volkswagen AG(大众汽车股份公司)

Paul Schwerdtner, Florens Greßner, Nikhil Kapoor, Felix Assion, René Sass, Wiebke Günther, Fabian Hüger, Peter Schlicht

摘要:

在本文中,我们提出了一个框架,用于评估在指定环境中部署机器学习模型的风险。为此,我们将风险的定义从决策理论转移到机器学习。我们开发并实现了一种方法,该方法允许定义部署场景,在每个场景中指定的条件下测试机器学习模型,并估计与测试中的机器学习模型的输出相关的损坏。使用每个场景的可能性以及估计的损害,我们定义了机器学习模型的\emph{key risk indicators}。场景的定义及其可能性的加权允许在多个应用领域的机器学习中进行标准化的风险评估。特别是,在我们的框架中,可以评估机器学习模型对随机输入损坏、由变化的环境引起的分布偏移和对抗性扰动的鲁棒性。

英文摘要:

In this paper we propose a framework for assessing the risk associated with deploying a machine learning model in a specified environment. For that we carry over the risk definition from decision theory to machine learning. We develop and implement a method that allows to define deployment scenarios, test the machine learning model under the conditions specified in each scenario, and estimate the damage associated with the output of the machine learning model under test. Using the likelihood of each scenario together with the estimated damage we define \emph{key risk indicators} of a machine learning model. The definition of scenarios and weighting by their likelihood allows for standardized risk assessment in machine learning throughout multiple domains of application. In particular, in our framework, the robustness of a machine learning model to random input corruptions, distributional shifts caused by a changing environment, and adversarial perturbations can be assessed.

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。

↑