知识卡片
数据是负债
数据既是 ML 的资产,也可能成为负债。采集、收集与治理数据会暴露出预期之外的细微复杂性与风险,处理不当会伤害用户与组织;医疗记录与求职记录所需的处理方式就截然不同。高成本才是把好数据“策展”出来的代价,没有免费午餐。收集与否不是技术问题而是政策与治理问题(哪些 PII 可收集、如何授权与删除)。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
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
- Data Management Principles · 查看原书依据
Whether and how to collect data is not a technical question. It is a question of policy and governance. (Parts of this topic are covered much more thoroughly in Chapter 6.) If we are allowed to collect and store the data, it must be secured from external access.
参考来源
- 位置:《Reliable Machine Learning》第02章《Data Management Principles》(源文件:extracted/Reliable Machine Learning/_epub-src/OEBPS/ch02.xhtml)
- 结论依据:原文明确“more data == better”是误区,强调数据在错误情形下是负债,且合规是政策治理而非技术问题。
- 原始内容:Whether and how to collect data is not a technical question. It is a question of policy and governance. (Parts of this topic are covered much more thoroughly in Chapter 6.) If we are allowed to collect and store the data, it must be secured from external access.