Structured Neural Summarization

  • Patrick Fernandes ,
  • Miltos Allamanis ,
  • Marc Brockschmidt

2018 International Conference on Learning Representations |

Publication | Publication | Publication

Summarization of long sequences into a concise statement is a core problem in natural language processing, requiring non-trivial understanding of the input. Based on the promising results of graph neural networks on highly structured data, we develop a framework to extend existing sequence encoders with a graph component that can reason about long-distance relationships in weakly structured data such as text. In an extensive evaluation, we show that the resulting hybrid sequence-graph models outperform both pure sequence models as well as pure graph models on a range of summarization tasks.