知识卡片
优化损失与泛化承担不同目标
优化让训练目标下降,泛化考察未参与拟合的新样本表现。更低训练损失不能单独证明未来误差更小;模型、数据质量、分布变化与评价协议都可能影响两者关系。
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
- · 查看原书依据
In many applications, the ML problems are formulated to minimize some loss function where the goal is the minimization of some loss function on a training set of data. Loss functions indicate the difference or the discrepancy between the predictions of the model being trained and the actual problem instances. For example, in classification problems, a label is assigned to the instances and models are
参考来源
- 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第03章《机器学习流程与数据管理》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/14_chapter03.xhtml)
- 结论依据:摘录中关于优化损失与泛化承担不同目标的定义、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:In many applications, the ML problems are formulated to minimize some loss function where the goal is the minimization of some loss function on a training set of data. Loss functions indicate the difference or the discrepancy between the predictions of the model being trained and the actual problem instances. For example, in classification problems, a label is assigned to the instances and models are