Discourse-Aware Neural Rewards for Coherent Text Generation

@inproceedings{Bosselut2018DiscourseAwareNR,
  title={Discourse-Aware Neural Rewards for Coherent Text Generation},
  author={Antoine Bosselut and A. Çelikyilmaz and X. He and Jianfeng Gao and Po-Sen Huang and Yejin Choi},
  booktitle={NAACL-HLT},
  year={2018}
}
  • Antoine Bosselut, A. Çelikyilmaz, +3 authors Yejin Choi
  • Published in NAACL-HLT 2018
  • Computer Science
  • In this paper, we investigate the use of discourse-aware rewards with reinforcement learning to guide a model to generate long, coherent text. In particular, we propose to learn neural rewards to model cross-sentence ordering as a means to approximate desired discourse structure. Empirical results demonstrate that a generator trained with the learned reward produces more coherent and less repetitive text than models trained with cross-entropy or with reinforcement learning with commonly used… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Explore Further: Topics Discussed in This Paper

    Modeling Coherence for Discourse Neural Machine Translation
    • 32
    • Highly Influenced
    • PDF
    Towards Coherent and Cohesive Long-form Text Generation
    • 9
    • PDF
    Evaluation Benchmarks and Learning Criteriafor Discourse-Aware Sentence Representations
    • 7
    • PDF
    Learning to Generalize from Sparse and Underspecified Rewards
    • 31
    • PDF

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 42 REFERENCES
    A Neural Attention Model for Abstractive Sentence Summarization
    • 1,516
    • Highly Influential
    • PDF
    Globally Coherent Text Generation with Neural Checklist Models
    • 124
    • PDF
    Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
    • 9,484
    • PDF
    A Deep Reinforced Model for Abstractive Summarization
    • 672
    • Highly Influential
    • PDF
    Deep Reinforcement Learning for Dialogue Generation
    • 715
    • PDF
    Challenges in Data-to-Document Generation
    • 176
    • PDF
    Paraphrase Generation with Deep Reinforcement Learning
    • 96
    • PDF
    Catching the Drift: Probabilistic Content Models, with Applications to Generation and Summarization
    • 355
    • PDF
    A Model of Coherence Based on Distributed Sentence Representation
    • 90
    • PDF
    Reinforced Video Captioning with Entailment Rewards
    • 63
    • Highly Influential
    • PDF