目录 概 主要内容 Pang T, Du C, Zhu J, et al. Max-Mahalanobis Linear Discriminant Analysis Networks[C]. international conference on machine learning, 2018: 4013-4022. @article{pang2018max-mahalanobis, title={Max-Mahalanobis Linear Discriminant Analysis Netwo
Suppose that we model each class density as multivariate Gaussian, in practice we do not know the parameters of the Gaussian distributions, and will need to estimate them using our training data: πk=Nk/N, where Nk is the number of class-k observation
最近跑深度学习,提出的feature是4096维的,放到我们的程序里,跑得很慢,很慢.... 于是,一怒之下,就给他降维处理了,但是matlab 自带的什么pca( ), princomp( )函数,搞不清楚怎么用的,表示不大明白,下了一个软件包: 名字:Matlab Toolbox for Dimensionality Reduction 链接:http://lvdmaaten.github.io/drtoolbox/ Currently, the Matlab Toolbox for Dim
修正线性单元(Rectified linear unit,ReLU) Rectified linear unit 在神经网络中,常用到的激活函数有sigmoid函数f(x)=11+exp(−x).双曲正切函数f(x)=tanh(x),今天要说的是另外一种activation function,rectified linear function,f(x)=max(0,x), The rectifier is, as of 2015, the most popular activation func