/DeepMatching

Classic deep neural network models for text matching, and implementation with tensorflow.

Primary LanguagePythonApache License 2.0Apache-2.0

DeepMatching

This repo is a collection for some classic deep neural network models in text matching , e.g. DSSM、L2R、MV-LSTM、MatchPyramid、Self-Attention、DSA...

Requirements

  • python 3.6
  • tensorflow 1.3.0

References

  1. DSSM: Learning deep structured semantic models for web search using clickthrough data, PS Huang et al.
  2. ARC: Convolutional Neural Network Architectures for Matching Natural Language Sentences, B Hu et al.
  3. MV-LSTM: A Deep Architecture for Semantic Matching with Multiple Positional Sentence Representations, S Wan et al
  4. MatchPyramid: Text Matching as Image Recognition, L Pang et al.
  5. Self-Attention: A Structured Self-attentive Sentence Embedding, Z Lin et al.
  6. BiMPM: Bilateral Multi-Perspective Matching for Natural Language Sentences, Z Wang et al.
  7. DIIN: Natural Language Infefence Over Interaction space, Y Gong et al.
  8. DRCN: Semantic Sentence Matching with Densely-connected Recurrent and Co-attentive Information, S Kim et al.
  9. DSA: Dynamic Self-Attention: Computing Attention over Words Dynamically for Sentence Embedding