Estimating the Brittleness of AI: Safety Integrity Levels and the Need for Testing Out-Of-Distributi

link之家

链接快照平台

输入网页链接，自动生成快照
标签化管理网页链接

相关文章推荐

气势凌人的柚子 · Numpy，Opencv及PIL.Image ...· 1 年前 ·

英姿勃勃的菠萝 · python生成点云数据_mob649e81 ...· 1 年前 ·

低调的回锅肉 · gdb调试进阶教程_12730336的技术博 ...· 1 年前 ·

慷慨的枕头 · node-sass 安装失败 ...· 1 年前 ·

发呆的单车 · 记录file_get_contents返回空 ...· 1 年前 ·

Test, Evaluation, Verification, and Validation (TEVV) for Artificial Intelligence (AI) is a challenge that threatens to limit the economic and societal rewards that AI researchers have devoted themselves to producing. A central task of TEVV for AI is estimating brittleness, where brittleness implies that the system functions well within some bounds and poorly outside of those bounds. This paper argues that neither of those criteria are certain of Deep Neural Networks. First, highly touted AI successes (eg. image classification and speech recognition) are orders of magnitude more failure-prone than are typically certified in critical systems even within design bounds (perfectly in-distribution sampling). Second, performance falls off only gradually as inputs become further Out-Of-Distribution (OOD). Enhanced emphasis is needed on designing systems that are resilient despite failure-prone AI components as well as on evaluating and improving OOD performance in order to get AI to where it can clear the challenging hurdles of TEVV and certification. 中文翻译：

推荐文章

气势凌人的柚子 · Numpy，Opencv及PIL.Image 格式相互转换_opencv numpy 转 image-CSDN博客

1 年前

英姿勃勃的菠萝 · python生成点云数据_mob649e8161c39d的技术博客_51CTO博客

1 年前

低调的回锅肉 · gdb调试进阶教程_12730336的技术博客_51CTO博客

1 年前

慷慨的枕头 · node-sass 安装失败 win32-x64-57_binding.node - 疯子110 - 博客园

1 年前

发呆的单车 · 记录file_get_contents返回空字符串的问题-腾讯云开发者社区-腾讯云

1 年前