Supervised learning - Wikipedia
5
公开标注数
5
参与人数
2026-07-15 10:02:58
首次 Whisper
本页的公开 Whisper
划选高亮2026-07-15 13:23:58
原文高亮摘录
“There is no single learning algorithm that works best on all supervised learning problems”
Whisper 随想笔记
I mean, that's just the free lunch theorem, kind of obvious though.
划选高亮2026-07-15 13:14:58
原文高亮摘录
“There is no single learning algorithm that works best on all supervised learning problems”
Whisper 随想笔记
Truer words were never said, every dataset's a different beast.
划选高亮2026-07-15 10:20:58
原文高亮摘录
“The goal of supervised learning is for the trained model to accurately predict the output for new, unseen data.”
Whisper 随想笔记
I've seen models nail the training set then choke on new data. Classic overfit.
划选高亮2026-07-15 10:11:58
原文高亮摘录
“The goal of supervised learning is for the trained model to accurately predict the output for new, unseen data.”
Whisper 随想笔记
Generalization error—that's the real boss, not just fitting the training set.
划选高亮2026-07-15 10:02:58
原文高亮摘录
“The goal of supervised learning is for the trained model to accurately predict the output for new, unseen data.”
Whisper 随想笔记
Yeah, but if the training data's biased, that 'unseen' data won't be so accurate.
分享本页 Whisper
短链接
https://domwhisper.com/s/6147d28e7f07嵌入代码
<iframe src="https://domwhisper.com/embed/6147d28e7f07" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>