Retrieval-Enhanced Adversarial Training for Neural Response Generation

  title={Retrieval-Enhanced Adversarial Training for Neural Response Generation},
  author={Qingfu Zhu and Lei Cui and W. Zhang and Furu Wei and Yi-ning Chen and T. Liu},
  • Qingfu Zhu, Lei Cui, +3 authors T. Liu
  • Published in ACL 2019
  • Computer Science
  • Dialogue systems are usually built on either generation-based or retrieval-based approaches, yet they do not benefit from the advantages of different models. [...] Key Method Distinct from existing approaches, the REAT method leverages an encoder-decoder framework in terms of an adversarial training paradigm, while taking advantage of N-best response candidates from a retrieval-based system to construct the discriminator. An empirical study on a large scale public available benchmark dataset shows that the REAT…Expand Abstract
    16 Citations

    Figures, Tables, and Topics from this paper

    Learning to Discriminate Perturbations for Blocking Adversarial Attacks in Text Classification
    • 15
    • PDF
    Cross Copy Network for Dialogue Generation
    • PDF
    Prototype-to-Style: Dialogue Generation with Style-Aware Editing on Retrieval Memory
    • 1
    • PDF
    Retrieval-Augmented Controllable Review Generation
    • PDF
    Condition-Transforming Variational Autoencoder for Generating Diverse Short Text Conversations
    Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
    • 23
    • PDF


    Adversarial Learning for Neural Dialogue Generation
    • 623
    • Highly Influential
    • PDF
    Adversarial Ranking for Language Generation
    • 188
    • PDF
    Generating Informative and Diverse Conversational Responses via Adversarial Information Maximization
    • 124
    • PDF
    A Diversity-Promoting Objective Function for Neural Conversation Models
    • 977
    • Highly Influential
    • PDF
    SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
    • 1,241
    • PDF
    An Ensemble of Retrieval-Based and Generation-Based Human-Computer Conversation Systems
    • 54
    • Highly Influential
    • PDF
    A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues
    • 698
    • PDF