BASS: Boosting Abstractive Summarization with Unified Semantic Graph
This work addresses a bottleneck in NLP for applications requiring concise summaries of extensive texts, representing an incremental advancement over existing methods.
The paper tackles the challenge of abstractive summarization for long or multi-documents by proposing BASS, a framework that uses a unified semantic graph to aggregate long-distance relations, resulting in substantial improvements in performance for these tasks.
Abstractive summarization for long-document or multi-document remains challenging for the Seq2Seq architecture, as Seq2Seq is not good at analyzing long-distance relations in text. In this paper, we present BASS, a novel framework for Boosting Abstractive Summarization based on a unified Semantic graph, which aggregates co-referent phrases distributing across a long range of context and conveys rich relations between phrases. Further, a graph-based encoder-decoder model is proposed to improve both the document representation and summary generation process by leveraging the graph structure. Specifically, several graph augmentation methods are designed to encode both the explicit and implicit relations in the text while the graph-propagation attention mechanism is developed in the decoder to select salient content into the summary. Empirical results show that the proposed architecture brings substantial improvements for both long-document and multi-document summarization tasks.