论文 – 有组织在! https://uzzz.org/ Mon, 08 Jul 2019 02:55:43 +0000 en-US hourly 1 https://wordpress.org/?v=5.2.4 https://uzzz.org/wp-content/uploads/2019/10/cropped-icon-32x32.png 论文 – 有组织在! https://uzzz.org/ 32 32 Learning sparse network using target dropout(文末有代码链接) https://uzzz.org/article/2322/ Mon, 08 Jul 2019 02:55:43 +0000 https://uzzz.org/article/2322.html 摘要
当神经网络的权值的数量超过了从输入映射到输出需要的权值数量,神经网络会更容易优化。这里暗存了一个两个阶段的学习进程:首先学习一个大的网络,然后删除连接或隐藏的单元。但是,标准的训练并不一定会使得网络易于修剪。于是,我们介绍了一种训练神经网络的方法——target…

The post Learning sparse network using target dropout(文末有代码链接) appeared first on 有组织在!.

]]>
The post Learning sparse network using target dropout(文末有代码链接) appeared first on 有组织在!.

]]>
Graph Convolutional Neural Networks for Web-Scale Recommender Systems(用于Web级推荐系统的图形卷积神经网络) https://uzzz.org/article/1588/ Sun, 07 Jul 2019 01:52:26 +0000 http://wp.uzzz.org/article/1588.html Graph Convolutional Neural Networks for Web-Scale Recommender Systems
用于Web级推荐系统的图形卷积神经网络
ABSTRACT
Recent advancements…

The post Graph Convolutional Neural Networks for Web-Scale Recommender Systems(用于Web级推荐系统的图形卷积神经网络) appeared first on 有组织在!.

]]>
The post Graph Convolutional Neural Networks for Web-Scale Recommender Systems(用于Web级推荐系统的图形卷积神经网络) appeared first on 有组织在!.

]]>
DCN:Deep & Cross Network for Ad Click Predictions简介 https://uzzz.org/article/1604/ Sat, 16 Mar 2019 08:16:39 +0000 http://wp.uzzz.org/article/1604.html Deep & Cross Network for Ad Click Predictions
摘要
作者起草了DCN,该网络可以保持DNN的优点(隐式地生成特征之间的交互),同时又利用交叉网络来对特征进行显式的交叉计算。这也不要求手工的特征工程,同时只是在DNN的基础上加了一些可容忍的复杂度。实验证明DCN已经在CTR预估与分类问题上超过了sota。…

The post DCN:Deep & Cross Network for Ad Click Predictions简介 appeared first on 有组织在!.

]]>
The post DCN:Deep & Cross Network for Ad Click Predictions简介 appeared first on 有组织在!.

]]>
《Learning Deep Structured Semantic Models for Web Search using Clickthrough Data 》论文总结 https://uzzz.org/article/1578/ Tue, 12 Mar 2019 14:14:10 +0000 http://wp.uzzz.org/article/1578.html 1.背景
DSSM是Deep Structured Semantic Model的缩写,即我们通常说的基于深度网络的语义模型,其核心思想是将query和doc映射到到共同维度的语义空间中,通过最大化query和doc语义向量之间的余弦相似度,从而训练得到隐含语义模型,达到检索的目的。DSSM有很广泛的应用,比如:搜索引擎检索,广告相关性,问答系统,机器翻译等。…

The post 《Learning Deep Structured Semantic Models for Web Search using Clickthrough Data 》论文总结 appeared first on 有组织在!.

]]>
The post 《Learning Deep Structured Semantic Models for Web Search using Clickthrough Data 》论文总结 appeared first on 有组织在!.

]]>
DehazeNet: An End-to-End System for Single Image Haze Removal https://uzzz.org/article/2994/ Fri, 04 May 2018 11:15:37 +0000 https://uzzz.org/article/2994.html 项目主页:http://caibolun.github.io/DehazeNet/
GitHub代码 :https://github.com/caibolun/DehazeNet
BReLU+Caffe :https://github.com/zlinker/mycaffe…

The post DehazeNet: An End-to-End System for Single Image Haze Removal appeared first on 有组织在!.

]]>
The post DehazeNet: An End-to-End System for Single Image Haze Removal appeared first on 有组织在!.

]]>
DeepCoNN https://uzzz.org/article/1717/ Thu, 15 Mar 2018 01:17:52 +0000 http://wp.uzzz.org/article/1717.html DeepCoNN Lei Zheng, Vahid Noroozi, and Philip S. Yu. 2017. Joint Deep Modeling of Users and Items Using Reviews for Recommendation.…

The post DeepCoNN appeared first on 有组织在!.

]]>
The post DeepCoNN appeared first on 有组织在!.

]]>
深度学习之图像修复 https://uzzz.org/article/2968/ Sun, 19 Mar 2017 10:25:44 +0000 https://uzzz.org/article/2968.html 图像修复问题就是还原图像中缺失的部分。基于图像中已有信息,去还原图像中的缺失部分。

从直观上看,这个问题能否解决是看情况的,还原的关键在于剩余信息的使用,剩余信息中如果存在有缺失部分信息的patch,那么剩下的问题就是从剩余信息中判断缺失部分与哪一部分相似。而这,就是现在比较流行的PatchMatch的基本思想。…

The post 深度学习之图像修复 appeared first on 有组织在!.

]]>
The post 深度学习之图像修复 appeared first on 有组织在!.

]]>
暗黑破坏神 2 私服 sf 114.215.178.67 https://uzzz.org/article/3196/ Wed, 09 Mar 2016 06:22:55 +0000 https://uzzz.org/article/3196.html 注册表
REGEDIT4
[HKEY_CURRENT_USER\Software\Blizzard Entertainment\Diablo II]

“BNETIP”=”114.215.178.67”…

The post 暗黑破坏神 2 私服 sf 114.215.178.67 appeared first on 有组织在!.

]]>
The post 暗黑破坏神 2 私服 sf 114.215.178.67 appeared first on 有组织在!.

]]>