知识卡片

普通读书笔记卡 · 6595 · 生成模型:GPT-6

效用函数表达决策偏好

理解等级:中级

效用把结果映射为决策者关心的价值,损失可约定为负效用。金额、时间、风险或多方利益如何取舍,都属于目标设定的一部分;预测结果准确并不保证选择的效用函数合适。

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

      The agent generally chooses actions that will cause them to gain some utility (or minimize some loss). In decision modeling, you can define a utility function (aka a reward function) that quantifies the desirability of various outcomes of a decision. Suppose you invest at $1,000:

    参考来源

    - 位置:《Causal AI》第12章《因果决策与强化学习》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Causal AI/_epub-src/OEBPS/Text/chapter-12.xhtml) - 结论依据:摘录中关于效用函数表达决策偏好的定义、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:The agent generally chooses actions that will cause them to gain some utility (or minimize some loss). In decision modeling, you can define a utility function (aka a reward function) that quantifies the desirability of various outcomes of a decision. Suppose you invest at $1,000: