← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8792.c · 生成模型:GPT-6

策略即代码连接原则与证据

理解等级:高级

把能明确定义的治理规则转为自动检查,如用途与同意记录、隐私检查、偏差指标和实验审计,使每次变化都留下可检查证据。机器只能验证被编码的条件,不能独自决定价值冲突或证明总体合规;难以自动化的判断仍需负责的人处理。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Defining and enforcing policy-as-code MLOps and CD4ML enable the automated application of Responsible AI policies at various points in the software development lifecycle. For instance, we may validate that training data has been obtained with consent for a particular purpose, and produce documentation and a journal of experiments. Bias tests or data privacy tests may be defined as a fitness function that is part of

    参考来源

    - 位置:《Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)》第09章《MLOps与持续交付》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)/_epub-src/OEBPS/ch09.html) - 结论依据:摘录中关于策略即代码连接原则与证据的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Defining and enforcing policy-as-code MLOps and CD4ML enable the automated application of Responsible AI policies at various points in the software development lifecycle. For instance, we may validate that training data has been obtained with consent for a particular purpose, and produce documentation and a journal of experiments. Bias tests or data privacy tests may be defined as a fitness function that is part of