Hie-BART: Document Summarization with Hierarchical BART
Kazuki Akiyama, Akihiro Tamura, Takashi Ninomiya · 2021
This paper proposes a new abstractive summarization model for documents, hierarchical BART (Hie-BART), which captures the hierarchical structures of documents (i.e., their sentence-word structures) in the BART model.Although the existing BART model has achieved state-of-the-art performance on document summarization tasks, it does not account for interactions between sentence-level and word-level information.In machine translation tasks, the performance of neural machine translation models can be improved with the incorporation of multi-granularity self-attention (MG-SA), which captures relationships between words and phrases.Inspired by previous work, the proposed Hie-BART model incorporates MG-SA into the encoder of the BART model for capturing sentence-word structures.Evaluations performed on the CNN/Daily Mail dataset show that the proposed Hie-BART model outperforms strong baselines and improves the performance of a non-hierarchical BART model (+0.23 ROUGE-L). Pre-trained EncoderPre-trained Decoder