← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 9384.c · 生成模型:GPT-6

增强变换要保留目标语义

理解等级:中级

增强通过扰动或生成增加变式,前提是变换后原标签仍成立。图像反射或旋转可能改变数字,文本同义替换或大小写处理也可能改变实体与意图;应核验有效性及与部署分布的差距。

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

      While it’s a little trickier, in the right circumstances, the technique of augmenting the data can be your friend. The concept behind augmentation is to generate novel data, either from perturbations of the existing data or by generating it from scratch. AffNIST2 is such an example. The famous MNIST dataset is a set of handwritten digits, 0 through 9 (see figure D.5). AffNIST takes each

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》附录D《附录D:机器学习工具与技术》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/appendix-d.html) - 结论依据:摘录中关于增强变换要保留目标语义的定义、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:While it’s a little trickier, in the right circumstances, the technique of augmenting the data can be your friend. The concept behind augmentation is to generate novel data, either from perturbations of the existing data or by generating it from scratch. AffNIST2 is such an example. The famous MNIST dataset is a set of handwritten digits, 0 through 9 (see figure D.5). AffNIST takes each