papering - 博客园

2017年9月27日

摘要： https://ipfs.io/ipfs/QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco/wiki/Convex_function.html Every norm is a convex function, by the triangle inequal 阅读全文

posted @ 2017-09-27 11:05 papering 阅读(245) 评论(0) 推荐(0)

Moore-Penrose Matrix Inverse 摩尔－彭若斯广义逆埃尔米特矩阵 Hermitian matrix

摘要： http://mathworld.wolfram.com/Moore-PenroseMatrixInverse.html 显然，埃尔米特矩阵主对角线上的元素都是实数的，其特征值也是实数。对于只包含实数元素的矩阵（实矩阵），如果它是对称阵，即所有元素关于主对角线对称，那么它也是埃尔米特矩阵。也就是说，阅读全文

posted @ 2017-09-27 09:14 papering 阅读(1158) 评论(0) 推荐(0)

2017年9月26日

But what exactly do we mean by "gets closer to"?

摘要： https://rdipietro.github.io/friendly-intro-to-cross-entropy-loss/ 【将输入转化为输出：概率分布】 When we develop a model for probabilistic classification, we aim to 阅读全文

posted @ 2017-09-26 21:00 papering 阅读(230) 评论(0) 推荐(0)

information entropy as a measure of the uncertainty in a message while essentially inventing the field of information theory

摘要： https://en.wikipedia.org/wiki/Claude_Shannon In 1948, the promised memorandum appeared as "A Mathematical Theory of Communication," an article in two 阅读全文

posted @ 2017-09-26 19:48 papering 阅读(175) 评论(0) 推荐(0)

SVM vs. Softmax

摘要： http://cs231n.github.io/linear-classify/ 阅读全文

posted @ 2017-09-26 19:23 papering 阅读(159) 评论(0) 推荐(0)

every row of W is a classifier for one of the classes

摘要： every row of W is a classifier for one of the classes As we saw above, every row of W is a classifier for one of the classes. The geometric interpreta 阅读全文

posted @ 2017-09-26 16:18 papering 阅读(214) 评论(0) 推荐(0)

Hinge Loss

摘要： The hinge loss is a convex function, so many of the usual convex optimizers used in machine learning can work with it. It is not differentiable, but h 阅读全文

posted @ 2017-09-26 15:42 papering 阅读(385) 评论(0) 推荐(0)

recognition rate generalization识别率泛化

摘要： http://www1.inf.tu-dresden.de/~ds24/lehre/ml_ws_2013/ml_11_hinge.pdf Two extremes: • Big 𝐶 → the loss is more important → better recognition rate but 阅读全文

posted @ 2017-09-26 15:35 papering 阅读(289) 评论(0) 推荐(0)

polynomial time 图灵机多项式时间

摘要：图灵的基本思想图灵的基本思想是用机器来模拟人们用纸笔进行数学运算的过程，他把这样的过程看作下列两种简单的动作：在纸上写上或擦除某个符号；把注意力从纸的一处移动到另一处；而在每个阶段，人要决定下一步的动作，依赖于（a）此人当前所关注的纸上某个位置的符号和（b）此人当前思维的状态。在可计算性理阅读全文

posted @ 2017-09-26 14:04 papering 阅读(891) 评论(0) 推荐(0)

Conditional random fields

摘要： crf 阅读全文

posted @ 2017-09-26 12:17 papering 阅读(124) 评论(0) 推荐(0)

Frobenius Norm

摘要： http://mathworld.wolfram.com/FrobeniusNorm.html 阅读全文

posted @ 2017-09-26 10:49 papering 阅读(190) 评论(0) 推荐(0)

L2 范数 L1 范数出租车范数

摘要： https://en.wikipedia.org/wiki/Norm_(mathematics) http://cs231n.github.io/classification/ 阅读全文

posted @ 2017-09-26 10:23 papering 阅读(249) 评论(0) 推荐(0)

2017年9月25日

卷积核即滤波器

摘要：输入输出的中间阅读全文

posted @ 2017-09-25 22:43 papering 阅读(194) 评论(0) 推荐(0)

Full卷积 Same卷积 Vaild卷积

摘要：阅读全文

posted @ 2017-09-25 22:03 papering 阅读(420) 评论(0) 推荐(0)

position

摘要： C:\Python36\python.exe D:/pymine/clean/chained_located/chained_located_dynamic_input.py '-69,-47,,,-72,-40,-37,-96,-36,-97,-67,-67,-43,,-100,-70,-54,-62,-92,-98,,-33,-77,-17,-17,,-98,-76... 阅读全文

posted @ 2017-09-25 18:13 papering 阅读(298) 评论(0) 推荐(0)

What happens when we continue stacking deeper layers on a “plain” convolutional neural network?

摘要： http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture9.pdf The deeper model performs worse, but it’s not caused by overfitting! 阅读全文

posted @ 2017-09-25 16:12 papering 阅读(282) 评论(0) 推荐(0)

A vectorized example

摘要： http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture4.pdf 阅读全文

posted @ 2017-09-25 14:19 papering 阅读(132) 评论(0) 推荐(0)

SVM vs. Softmax

摘要： http://cs231n.github.io/linear-classify/ http://cs231n.github.io/assets/svmvssoftmax.png 阅读全文

posted @ 2017-09-25 11:28 papering 阅读(198) 评论(0) 推荐(0)

2017年9月22日

run length encoding

摘要：直观阅读全文

posted @ 2017-09-22 16:16 papering 阅读(128) 评论(0) 推荐(0)

2017年9月20日

Data Decisions: DSP vs. DMP

摘要： http://www.cmo.com/features/articles/2016/3/9/data-decisions-dsp-vs-dmp.html As marketers assess their requirements for marketing technology, the ques 阅读全文

posted @ 2017-09-20 17:11 papering 阅读(275) 评论(0) 推荐(0)

What I learned from competing against a ConvNet on ImageNet

摘要： http://karpathy.github.io/2014/09/02/what-i-learned-from-competing-against-a-convnet-on-imagenet/ 阅读全文

posted @ 2017-09-20 17:01 papering 阅读(197) 评论(0) 推荐(0)

Image Recognition

摘要： https://www.tensorflow.org/tutorials/image_recognition 阅读全文

posted @ 2017-09-20 16:55 papering 阅读(229) 评论(0) 推荐(0)

They're much closer in spirit to how our brains work than feedforward networks.

摘要： http://neuralnetworksanddeeplearning.com/chap1.html Up to now, we've been discussing neural networks where the output from one layer is used as input 阅读全文

posted @ 2017-09-20 16:26 papering 阅读(247) 评论(0) 推荐(0)

output value . Sigmoid neurons are similar to perceptrons, but modified so that small changes in their weights and bias cause only a small change in their output.

摘要： http://neuralnetworksanddeeplearning.com/chap1.html . Sigmoid neurons are similar to perceptrons, but modified so that small changes in their weights 阅读全文

posted @ 2017-09-20 15:18 papering 阅读(296) 评论(0) 推荐(0)

2017年9月19日

aerospike

摘要： spike 阅读全文

posted @ 2017-09-19 20:33 papering 阅读(182) 评论(0) 推荐(0)

它是个黑盒，无法解释的黑盒；它避免了我们的或者某种来源的先验强加于陌生人：另外一个黑盒。

摘要： @Matthew 【[抱拳]】推荐系统中，传统的两大算法，无论是基于人的过滤，还是基于物品的过滤，在前期历史数据的量和质都不充足的情况下，这两种传统的算法都无法冷启动。【基于ANN的一种解决办法】张三第一次点击了汽车广告，通过ANN训练，使得输出层的汽车的概率最大，次之的就为接下来要投出的广告类阅读全文

posted @ 2017-09-19 20:28 papering 阅读(230) 评论(0) 推荐(0)

bilingual evaluation understudy

摘要： BLEU is designed to approximate human judgement at a corpus level, and performs badly if used to evaluate the quality of individual sentences. https:/ 阅读全文

posted @ 2017-09-19 18:36 papering 阅读(231) 评论(0) 推荐(0)

TensorFlow_action

摘要：安装TensorFlow 包依赖 https://www.tensorflow.org/tutorials/layers Building the CNN MNIST Classifier 【Each of these methods accepts a tensor as input and re 阅读全文

posted @ 2017-09-19 18:10 papering 阅读(518) 评论(0) 推荐(0)

Hadamard Product

摘要： http://mathworld.wolfram.com/HadamardProduct.html 阅读全文

posted @ 2017-09-19 16:21 papering 阅读(235) 评论(0) 推荐(0)

cross-correlation convolution. Hermitian_function

摘要： Hermitian_function 阅读全文

posted @ 2017-09-19 15:13 papering 阅读(131) 评论(0) 推荐(0)

S-形函数广泛应用于ANN 的激活函数

摘要： Logistic function hyperbolic tangent arctangent function Gudermannian function Error function Generalised logistic function Smoothstep function Specif 阅读全文

posted @ 2017-09-19 14:27 papering 阅读(615) 评论(0) 推荐(0)

减少75%的数据量它的精确位置远不及它和其他特征的相对位置的关系重要

摘要： zh.wikipedia.org/wiki/卷积神经网络池化（Pooling）是卷积神经网络中另一个重要的概念，它实际上一种形式的向下采样。有多种不同形式的非线性池化函数，而其中“最大池化（Max pooling）”是最为常见的。它是将输入的图像划分为若干个矩形区域，对每个子区域输出最大值。直觉上阅读全文

posted @ 2017-09-19 14:21 papering 阅读(283) 评论(0) 推荐(0)

The hidden layers are either convolutional, pooling or fully connected.

摘要： https://en.wikipedia.org/wiki/Convolutional_neural_network Convolutional layers apply a convolution operation to the input, passing the result to the 阅读全文

posted @ 2017-09-19 14:15 papering 阅读(236) 评论(0) 推荐(0)

images have the “stationarity” property, which implies that features that are useful in one region are also likely to be useful for other regions.

摘要： Convolutional networks may include local or global pooling layers[clarification needed], which combine the outputs of neuron clusters at one layer int 阅读全文

posted @ 2017-09-19 11:19 papering 阅读(217) 评论(0) 推荐(0)

2017年9月18日

算法报告

摘要：【要解决到的问题】“对于给定的mac-rssi列表，求出特征值。” 阅读全文

posted @ 2017-09-18 21:55 papering 阅读(415) 评论(0) 推荐(0)

word2vec_basic.py

摘要： https://raw.githubusercontent.com/tensorflow/tensorflow/r1.3/tensorflow/examples/tutorials/word2vec/word2vec_basic.py 阅读全文

posted @ 2017-09-18 20:08 papering 阅读(335) 评论(0) 推荐(0)

Softmax_function

摘要： en.wikipedia.org/wiki/Softmax_function 阅读全文

posted @ 2017-09-18 12:04 papering 阅读(154) 评论(0) 推荐(0)

DNN 数据的表示

摘要：阅读全文

posted @ 2017-09-18 04:31 papering 阅读(194) 评论(0) 推荐(0)

Convex combination

摘要： en.wikipedia.org/wiki/Convex_combination 凸组合 In convex geometry, a convex combination is a linear combination of points (which can be vectors, scalars 阅读全文

posted @ 2017-09-18 04:04 papering 阅读(796) 评论(0) 推荐(0)

好为什么好转变思维

摘要：阅读全文

posted @ 2017-09-18 03:41 papering 阅读(169) 评论(0) 推荐(0)