Structured Neural Summarization
This addresses the challenge of summarization in NLP, which is incremental as it builds on existing methods by integrating graph neural networks.
The authors tackled the problem of summarizing long sequences by extending sequence encoders with a graph component to reason about long-distance relationships in text, resulting in hybrid models that outperform pure sequence and graph models on summarization tasks.
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.