Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[regression,combinatorics,sampling]SparseRegression/DiscreteRegression:稀疏回归/离散回归,以及排列组合上的采样 #18

Open
shouldsee opened this issue Sep 6, 2019 · 1 comment

Comments

@shouldsee
Copy link
Contributor

shouldsee commented Sep 6, 2019

神经网络对离散性质的处理是糟糕的.尽管Softmax可以模拟(1,0,0)这种one-hot矢量,但是并不能完全满足(1,0,0) 中的后两位为零的约束条件.这就造成了一些信息的泄漏,导致模型约束不能满足.相比较之下,蒙特卡罗模拟能够更好地满足这种约束条件.但是MCMC/hill-climbing并不总是最速解

需要继续考察一下是否存在好的稀疏回归算法.

为了处理离散域上的采样,考虑如下分布.

$$
\begin{align}
x &\sim Perm(D,beta,S_n)\\
x &\in S_n \\
e.g.: \, x &= \{1,2,3,4,5\} \,or\, \{1,3,2,4,5\} \,or \dots \\
P(x) &= C \exp( - beta \sum_i D[i,x[i] ]  )
\end{align}
$$

目前没有找到现有的采样算法.望各路大神指教.我稍稍写了一个heuristics但是显然只能暂时用用

@shouldsee shouldsee changed the title [Stat]SparseRegression/DiscreteRegression:稀疏回归/离散回归 [regression,combinatorics,sampling]SparseRegression/DiscreteRegression:稀疏回归/离散回归,以及排列组合上的采样 Sep 8, 2019
@shouldsee
Copy link
Contributor Author

This looks remotely relevant
Equivariance Through Parameter-Sharing https://arxiv.org/abs/1702.08389

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant