标签:
长短期记忆网络LSTM(Long Short Term Memory)本身不是一个完整的模型,主要是对RNN隐含层的改进。因此,RNN网络即使用LSTM单元的RNN网络。LSTM非常适合用于处理与时间序列高度相关的问题,例如机器翻译、对话生成、编码解码、图文转换等。
说明:word2vec(CBOW,Skip-gram),GRU,word embedding(词向量),MDP(Markov Decision Process),Deep Reinforcement Learning,DQN等。
参考文献:
[1] LSTM实现详解:http://www.csdn.net/article/2015-09-14/2825693
[2] char-rnn:https://github.com/karpathy/char-rnn
[3] 深入浅出LSTM神经网络:http://www.csdn.net/article/2015-06-05/2824880
[4] Learning to read with recurrent neural networks:http://blog.terminal.com/recurrent-neural-networks-deep-net-optimization-lstm/
[5] 理解LSTM网络:http://www.jianshu.com/p/9dc9f41f0b29/
[6] 深度学习BP算法的推导(附加RNN,LSTM的推导说明):http://blog.csdn.net/zhuanshenweiliu/article/details/42267993
[7] LSTM Networks for Sentiment Analysis:http://deeplearning.net/tutorial/lstm.html
[8] 如何评价最近比较火的LSTM?:http://www.zhihu.com/question/27017697
[9] Long Short-Term Memory:Tutorial on LSTM Recurrent Networks:http://people.idsia.ch/~juergen/lstm/
[10] caffe-lstm:https://github.com/junhyukoh/caffe-lstm
[11] LSTM简介以及数学推导:http://blog.csdn.net/a635661820/article/details/45390671
[12] LSTM与情感分析:http://www.weixingon.com/s/lstm+%E6%83%85%E6%84%9F%E5%88%86%E6%9E%90
[13] 有哪些LSTM(Long Short Term Memory)和RNN(Recurrent)网络的教程?:http://www.zhihu.com/question/29411132?utm_source=top.caibaojian.com/47897
[14] 深度学习资料大全:http://www.cnblogs.com/charlotte77/p/5485438.html
[15] 近期风靡互联网的Deep Dream人工智能图像识别软件:http://www.ltaaa.com/bbs/thread-364424-1-1.html
[16] 深度学习:推动NLP领域发展的新引擎:http://www.iteye.com/news/31261
[17] 盘点8个最具启发意义的深度学习应用:http://synchuman.baijia.baidu.com/article/542746
[18] DQN从入门到放弃(DQN与增强学习):https://zhuanlan.zhihu.com/p/21262246
[19] 深度增强学习DRL专栏:http://blog.csdn.net/column/details/deeprl.html
标签:
原文地址:http://www.cnblogs.com/shengshengwang/p/5689152.html