-
论文:Multi-Perspective Sentence Similarity Modeling with Convolutional Neural Networks
项目:github
导读:Tensorflow实例-CNN处理句子相似度(MPCNN) -
论文:Character-level Convolutional Networks for Text Classification
项目:github
导读:基于字符的卷积神经网络实现文本分类(char-level CNN)-论文详解及tensorflow实现 -
文章:Understanding LSTM Networks
项目:github
导读:基于循环神经网络实现基于字符的语言模型(char-level RNN Language Model)-tensorflow实现 -
论文:Implementation of Hierarchical Attention Networks for Document Classification
项目:github
导读:Implementation of Hierarchical Attention Networks for Document Classification的讲解与Tensorflow实现 -
论文:Ask Me Anything: Dynamic Memory Networks for Natural Language Processing
项目:github
导读:Ask Me Anything: Dynamic Memory Networks for Natural Language Processing 阅读笔记及tensorflow实现
总体来说,难度由浅入深,适合初学者入门。大部分项目都会按照论文、项目、和导读三个部分,首先阅读论文,其次在github上寻找开源实现,如果其中有不懂的问题就在网上找解决办法,我想这也是大多数初学者适合的道路。在写这篇入门实例推荐时,我也会把实例中涉及的代码风格进行统一,以方便记忆。目前涉及的模型主要有:
- Deep Neural Network(DNN)
- Convolution Neural Network(CNN)
- Recurrent Neural Network(RNN)以及LSTM
- Recurrent Convolution Neural Network(RCNN)
- sequence2sequence(Attention)
- Hierarchical Attention Networks(HAN)
- Dynamic Memory Networks(DMN)
- EntityNetwork
[1] 李航NSR论文:深度学习NLP的现有优势与未来挑战
[2] NLP维基百科