Linguistic Versus Latent Relations for Modeling Coherent Flow in Paragraphs

@inproceedings{Kang2019LinguisticVL,
  title={Linguistic Versus Latent Relations for Modeling Coherent Flow in Paragraphs},
  author={Dongyeop Kang and H. Hayashi and A. Black and E. Hovy},
  booktitle={EMNLP/IJCNLP},
  year={2019}
}
  • Dongyeop Kang, H. Hayashi, +1 author E. Hovy
  • Published in EMNLP/IJCNLP 2019
  • Computer Science
  • Generating a long, coherent text such as a paragraph requires a high-level control of different levels of relations between sentences (e.g., tense, coreference). We call such a logical connection between sentences as a (paragraph) flow. In order to produce a coherent flow of text, we explore two forms of intersentential relations in a paragraph: one is a human-created linguistical relation that forms a structure (e.g., discourse tree) and the other is a relation from latent representation… CONTINUE READING
    7 Citations
    Plan ahead: Self-Supervised Text Planning for Paragraph Completion Task
    • 1
    • PDF
    INSET: Sentence Infilling with Inter-sentential Generative Pre-training
    • 1
    INSET: Sentence Infilling with INter-SEntential Transformer
    • 3
    • PDF
    Narrative Incoherence Detection
    • PDF
    A-BPS: Automatic Business Process Discovery Service using Ordered Neurons LSTM
    • X. Han, L. Hu, +4 authors Pengwei Hu
    • Computer Science
    • 2020 IEEE International Conference on Web Services (ICWS)
    • 2020

    References

    SHOWING 1-10 OF 41 REFERENCES
    A Hierarchical Neural Autoencoder for Paragraphs and Documents
    • 483
    • PDF
    Hierarchical Recurrent Neural Network for Document Modeling
    • 126
    • Highly Influential
    • PDF
    Variational Neural Discourse Relation Recognizer
    • 12
    • PDF
    A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues
    • 697
    • Highly Influential
    • PDF
    Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks
    • 2,026
    • PDF
    Representation Learning for Text-level Discourse Parsing
    • 166
    • PDF
    Glove: Global Vectors for Word Representation
    • 16,735
    • Highly Influential
    • PDF
    Discourse Strategies for Generating Natural-Language Text*
    • M KathleenR.
    • 2002
    • 113
    • PDF
    Describing a Knowledge Base
    • 18
    • PDF
    Pragmatics and Natural Language Generation
    • E. Hovy
    • Computer Science
    • Artif. Intell.
    • 1990
    • 165
    • PDF