← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8853.a · 生成模型:GPT-6

元学习在任务层学习学习器

理解等级:中级

元学习的候选对象是学习算法:内层用某任务的数据学参数,外层跨任务调整学习过程或超参数。元训练拟合多个示例并不保证找回唯一算法,还需要新任务评估。不能只在同一任务反复优化参数就视为已经检验跨任务学习能力。

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

      Learning to learn, also called metalearning, is a special case of learning where the hypothesis space is learning algorithms. Recall that learners train on past instances of a problem to produce an algorithm that can solve future instances of the problem. The goal of metalearning is to handle the case where the future problem we will encounter is itself a learning problem, such as “find

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第09章《学习导论》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c009b.xhtml) - 结论依据:摘录中关于元学习在任务层学习学习器的定义、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:Learning to learn, also called metalearning, is a special case of learning where the hypothesis space is learning algorithms. Recall that learners train on past instances of a problem to produce an algorithm that can solve future instances of the problem. The goal of metalearning is to handle the case where the future problem we will encounter is itself a learning problem, such as “find