issues Search Results · repo:nndl/nndl.github.io language:HTML
Filter by
629 results
(46 ms)629 results
innndl/nndl.github.io (press backspace or delete to remove)本人为神经网络小白,在学习邱老师的《神经网络与深度学习》课程以及对应的资料书时,发现课程和书中都没有关于梯度的解释,以及为什么要做梯度下降,为什么解决梯度消失和梯度爆炸问题,虽然网上也有一些解释,但是希望邱老师能在课程和书中也能解释一下。
SandDandelion
- 3
- Opened on Aug 15, 2024
- #642
image
如图,递归神经网络( Recursive Neural Network)应该缩写为RvNNs,而不是RecNN。 RecNN是:https://github.com/awarebayes/RecNN (Reinforcement
Learning) 互联网上找不到关于RecNN是Recursive NN的说明
FeijiangHan
- Opened on Jun 16, 2024
- #640
p93 Borel可测函数定义错误
p93 侧边 “定义在实数空间 $\mathbb R^D$ 中的有界闭集上的任意连续函数,也称为 Borel 可测函数.” 显然这里对 Borel 可测函数的定义是错误的,可测性的定义与连续无关。
此处可能原本想说明别的东西,希望作者加以检查。
CSMsamuel
- 1
- Opened on Mar 12, 2024
- #638
问题描述(例如:错别字) + 页号或公式号
- [ ] 错别字:P66 公式2.3 下面一行“1111”改为“2222” (参考示例1)
- [ ] 推导错误:P66 公式2.3 中推导错误,。。。 (参考示例2)
硕士论文中使用文中图片是否需要说明呢
reatiny
- 2
- Opened on Jan 4, 2024
- #637
课后习题答案在哪里可以找到
- [ ] 老师,我想要课后习题答案(Teacher, how do I ask for the answers to the post-class exercises )
DoYangTan
- 3
- Opened on Jan 2, 2024
- #636
请教:课本P10及PPT绪论P32相关内容:一个好的表示学习策略必须具备一定深度,这个该怎么理解呢?
- [ ] 深度指原始数据进行非线性转换的次数,表示学习是指从底层特征抽取高层语义的过程。
- [ ] 具备一定深度的表示学习策略是因为给足隐藏层和隐藏单元理论上可以表示特征所有的抽象的组合吗?
- [ ] 所以是一个好的表示策略吗?
- [ ] (这个问题理解上感觉一直很模糊,非常希望大佬们帮忙解答一下,万分感谢!) ...
647615
- 2
- Opened on Dec 15, 2023
- #635

Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip!
Restrict your search to the title by using the in:title qualifier.
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip!
Restrict your search to the title by using the in:title qualifier.