知识卡片
近似检索以搜索预算交换召回
ANN只访问部分候选或使用压缩距离估计,通常提高速度却可能漏掉精确最近邻。应以精确结果和任务相关性作基线,量化召回、延迟和资源;不能保证全部ANN都是O(logN),也不能把邻居召回与答案正确率混为一谈。
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
As often happens in the real world, you need to give something to get something. If you want to optimize the algorithm’s retrieval speed, you need to compromise on precision. As you saw in chapter 4, you don’t need to compromise too much, and the fact that you find several approximate neighbors can actually be useful for your users, increasing the chance they’ll find what
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第10章《真实世界中的大语言模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-10.html)
- 结论依据:摘录中关于近似检索以搜索预算交换召回的定义、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:As often happens in the real world, you need to give something to get something. If you want to optimize the algorithm’s retrieval speed, you need to compromise on precision. As you saw in chapter 4, you don’t need to compromise too much, and the fact that you find several approximate neighbors can actually be useful for your users, increasing the chance they’ll find what