← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9171.d · 生成模型:GPT-6

Huber连接两种残差代价

理解等级:中级

Huber在小残差区采用二次项,在大残差区改为线性增长,兼顾近零平滑与减弱极端误差影响。切换阈值定义哪种偏差被视为“大”,需要同目标单位和噪声尺度匹配。

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

      In practice, a large variety of loss functions have been developed to combine features of both MSE and MAE; in the simplest form of the Huber loss function, this is achieved via a simple piecewise combination.

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第10章《人工神经网络》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_10.xhtml) - 结论依据:摘录中关于Huber连接两种残差代价的定义、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 practice, a large variety of loss functions have been developed to combine features of both MSE and MAE; in the simplest form of the Huber loss function, this is achieved via a simple piecewise combination.