技术笔记本

在有监督的机器学习中，经常会说到训练集（train)、验证集（validation）和测试集（test），这三个集合的区分可能会让人糊涂，特别是，有些读者搞不清楚验证集和测试集有什么区别。划分 #如果我们自己已经有了一个大的标注数据集，想要完成一个有监督模型的测试，那么通常使用均匀随机抽样的方式，将数据集划分为训练集、验证集、测试集，这三个集合不能有交集，常见的比例是8:1:1，当然比例是...

The Unreasonable Effectiveness of Recurrent Neural NetworksUnderstanding LSTM NetworksRECURRENT NEURAL NETWORKREGULARIZATIONNeural Networks and Deep Learning

LpL_p距离, L1L_1范数, 和 L2L_2范数LpL_p 距离定义为: Lp(xi,xj)=(∑l=1n|x(l)i−x(l)j|p)1p L_p(x_i, x_j) = \left( \sum_{l=1}^{n} | x_i^{(l)} - x_j^{(l)} |^p \right)^{\frac{1}{p}} 其中xi∈Rnx_i \in \mathbf{R}^n,

>>> import tensorflow as tfTraceback (most recent call last):  File "", line 1, in   File "tensorflow/__init__.py", line 23, in     from tensorflow.python import *  File "tensorflow/python/_

Machine Learning Group at Department of Engineering, University of CambridgeMachine Learning Group at University of columbiaMachine Learning Research Group at university of texas at austinMachin

Shannon entropy is one of the most important metrics in information theory. Entropy measures the uncertainty associated with a random variable, i.e. the expected value of the information in the messag

http://www.encyclopediaofmath.org/index.php/Bounded_operator

In linear algebra, a symmetric n × n real matrix M is said to be positive definite if zTMz is positive for every non-zero columnvector z of n real numbers. Here zT denotes thetranspose

DefinitionA Hilbert space H is a real or complex inner product space that is also a complete metric space with respect to the distance function induced by the inner product.[2] To say that

深度学习中的各种卷积网络大家知多少？对于那些听说过却又对它们没有特别清晰的认识的小伙伴们，Kunlun Bai这篇文章非常值得一读。Kunlun Bai 是一位人工智能、机器学习、物体学以及工程学领域的研究型科学家，在本文中，他详细地介绍了 2D、3D、1x1 、转置 、空洞（扩张）、空间可分离、深度可分离、扁平化、 分组等十多种卷积网络类型。AI 科技评论编译如下。如果你曾听过深度学习的各...