关闭

[置顶] 深度学习必备手册

摘要:深度学习作为人工智能的前沿技术,虽然一方面推动者人工智能的发展;但是人类的终极目标四强人工智能。最近也有一些关于类似于笔者认为的广度学习的出现,但是宗其所属,还是应该在深度学习发展的历史上前进。        深度学习的概念源于人工神经网络的研究,如果追溯深度学习的概念还是要回到2006年Hinton那篇论文,基于深信度网(DNB)提出非监督贪心逐层训练算法,未解决...
阅读(9) 评论(0)

[置顶] The Definitive Security Data Science and Machine Learning Guide

This is the Definitive Security Data Science and Machine Learning Guide. It includes books, tutorials, presentations, blog posts, and research papers about solving security problems using data science...
阅读(258) 评论(0)

[置顶] Deep Learning Online Courses/Tutorials 深度学习 资料

Stanford c231n Youtube MIT 6.S094: Introduction to Deep Learning and Self-Driving Car Oxford Machine Learning: 2014-2015 MIT 6.034 Artificial Intelligence, Fall 2010...
阅读(294) 评论(0)

[置顶] Omnispace 收藏夹

Android 安全 看雪安全论坛 乌云 吾爱破解 DroidSec安卓安全中文站 阿里聚安全 360安全播报平台 腾讯安全应急响应中心 360移动安全 wiki.secmobi.com mobilesecuritywiki.com android-security-awesome [C...
阅读(511) 评论(0)

BP,RNN 和 LSTM暨《Supervised Sequence Labelling with Recurrent Neural Networks-2012》阅读笔记

一、BackPropagation wljkwjkl:表示第l−1l−1层第k个神经元到第ll层第j个神经元的连接权重;bljbjl:表示第ll层第j个神经元的偏置;zljzjl:表示第ll层第j个神经元的带权输入;aljajl:表示第ll层第j个神经元的激活值;σσ:表示一个激活函数(sigmoid,relu,tanh); zlj=∑kwljkal−1k+bljzjl=∑kwjkla...
阅读(3) 评论(0)

Tensorflow中GRU和LSTM的权重初始化

GRU和LSTM权重初始化 在编写模型的时候,有时候你希望RNN用某种特别的方式初始化RNN的权重矩阵,比如xaiver或者orthogonal,这时候呢,只需要: 12345678910 cell = LSTMCell if self.args.use_lstm else GRUCellwith tf.variable_scope(initializer=tf....
阅读(4) 评论(0)

反卷积在神经网络可视化上的成功应用

反卷积(Deconvolution)的概念第一次出现是Zeiler在2010年发表的论文Deconvolutional networks中,但是并没有指定反卷积这个名字,反卷积这个术语正式的使用是在其之后的工作中(Adaptive deconvolutional networks for mid and high level feature learning)。随着反卷积在神经网络可视化上...
阅读(17) 评论(0)

Cython代码和Python代码区别

代码运行在IPython-Notebook中,在IPython-Notebook中导入cython环境。 1 %load_ext cython Cython可以在Python中掺杂C和C++的静态类型,cython编译器可以把Cython源码编译成C或C++代码,编译后的代码可以单独执行或者作为Python中的模型使用。Cython中的强大之处在于...
阅读(9) 评论(0)

卷积神经网络中不同类型的卷积方式介绍

之前在文章《牛逼的深度卷积神经网络CNN》详细介绍了卷积神经网络的基本原理,以及常见的基本模型,如LeNet,VGGNet,AlexNet,ReseNet,Inception Net的基本结构和原理。今天主要总结一下,卷积神经网络家族中,不同类型的卷积方式以及它们各自的优点。为了简单起见,我们仅从2维的角度介绍。 文末提供相关资料下载地址 卷积基本概念 首先,我们首先回顾一下卷积相关...
阅读(23) 评论(0)

TensorFlow 使用例子-LSTM实现序列标注

本文主要改写了一下"Sequence Tagging with Tensorflow"程序。原文是基于英文的命名实体识别(named entity recognition)问题,由于博主找不到相应的中文数据集(其实是没备份数据丢了,如果有同学提供,万分感谢)。因此,本文用了msra的分词数据。另外,由于用到了词向量,所以用了搜狗实验室发布的2008新闻数据,提前训练了300维度的字向量(用的gen...
阅读(184) 评论(0)

SimGAN-Captcha代码阅读与复现

项目介绍 项目地址:戳这里 大概的讲一下这个项目的起因是大神要参加HackMIT,需要他们在15000张验证码中识别出10000张或者每个字符的识别准确率要到90%。然后他不想标注数据(就是这么任性~)。于是决定先自己生成一批验证码(synthesizer合成器),然后把这些验证码用一个refiner(GAN)去对这批合成的验证码做一些调整让它们看起来和真实的训练样本的样式差不多。这样他就相当...
阅读(17) 评论(0)

Deep learning:一(基础知识_1)

前言:   最近打算稍微系统的学习下deep learing的一些理论知识,打算采用Andrew Ng的网页教程UFLDL Tutorial,据说这个教程写得浅显易懂,也不太长。不过在这这之前还是复习下machine learning的基础知识,见网页:http://openclassroom.stanford.edu/MainFolder/CoursePage.php?course=DeepL...
阅读(24) 评论(0)

CS231n Convolutional Neural Networks for Visual Recognition

Table of Contents: Architecture Overview ConvNet Layers Convolutional Layer Pooling Layer Normalization Layer Fully-Connected Layer Converting Fully-Connected Layers to Convolutional Layers...
阅读(23) 评论(0)

TensorFlow for Hackers (Part VII) - Credit Card Fraud Detection using Autoencoders in Keras

It’s Sunday morning, it’s quiet and you wake up with a big smile on your face. Today is going to be a great day! Except, your phone rings, rather “internationally”. You pick it up slowly and hear some...
阅读(54) 评论(0)

TensorFlow for Hackers (Part VI) - Human Activity Recognition using LSTMs on Android

Ever wondered how your smartphone, smartwatch or wristband knows when you’re walking, running or sitting? Well, your device probably has multiple sensors that give various information. GPS, audio (i.e...
阅读(48) 评论(0)

TensorFlow for Hackers (Part IV) - Neural Network from Scratch

Developing models using TensorFlow is easy and fun, but real understanding can be achieved only via reading and implementing the algorithms on your own. This time we will skip TensorFlow entirely and ...
阅读(40) 评论(0)

TensorFlow for Hackers - Part III

Have you ever stood still, contemplating about how cool would it be to build a model that can distinguish cats from dogs? Don’t be shy now! Of course you did! Let’s get going! Cat or a dog? We hav...
阅读(37) 评论(0)

TensorFlow for Hackers - Part II

In this one, you will learn how to create a Neural Network (NN) and use it for deciding whether a student has alcohol consumption problems. Do students drink too much? How can you predict that? What ...
阅读(34) 评论(0)

TensorFlow for Hackers - Part I

What is TensorFlow? TensorFlow is a library for number crunching created and maintained by Google. It’s used mainly for machine learning (especially deep learning) tasks. While still in beta (version...
阅读(40) 评论(0)

Implementing a CNN for Human Activity Recognition in Tensorflow

In the recent years, we have seen a rapid increase in smartphones usage which are equipped with sophisticated sensors such as accelerometer and gyroscope etc. These devices provide the opportunity for...
阅读(33) 评论(0)

RUNNING JUPYTER NOTEBOOKS ON A REMOTE SERVER VIA SSH

My roommate, Monica, introduced me to Jupyter Notebooks last year. And since then, I’ve been addicted to them for any Python coding I do, so much that I’m considering recreating this entire blog in Ju...
阅读(20) 评论(0)

RNNs in Tensorflow, a Practical Guide and Undocumented Features

In a previous tutorial series I went over some of the theory behind Recurrent Neural Networks (RNNs) and the implementation of a simple RNN from scratch. That’s a useful exercise, but in practice we...
阅读(18) 评论(0)

Learning Reinforcement Learning (with Code, Exercises and Solutions)

Why Study Reinforcement Learning Reinforcement Learning is one of the fields I’m most excited about. Over the past few years amazing results like learning to play Atari Games from raw pixelsand M...
阅读(16) 评论(0)
743条 共38页1 2 3 4 5 ... 下一页 尾页
    个人资料
    • 访问:313674次
    • 积分:4400
    • 等级:
    • 排名:第7188名
    • 原创:11篇
    • 转载:732篇
    • 译文:0篇
    • 评论:14条
    文章分类
    最新评论