Home > Research > Publications & Outputs > Joint abstractive and extractive method for lon...


View graph of relations

Joint abstractive and extractive method for long financial documentsummarization

Research output: Contribution to conference - Without ISBN/ISSN Conference paperpeer-review

Publication date26/10/2021
Number of pages7
<mark>Original language</mark>English
Event3rd Financial Narrative Processing Workshop (FNP 2021) - Virtual
Duration: 15/09/202116/09/2021


Workshop3rd Financial Narrative Processing Workshop (FNP 2021)
Internet address


In this paper we show the results of our participation in the FNS 2021 shared task. In our work we propose an end to end financial narrative summarization system that first selects salient sentences from the document and then
paraphrases extracted sentences. This method generates an overall concise summary that maximises the ROUGE metric with the gold standard summary. The end to end system is developed using a hybrid extractive and abstractive architecture followed by joint training using policy-based reinforcement learning to bridge together the two networks. Empirically, we achieve better scores than the proposed baselines and toplines of FNS 2021 (LexRank, TextRank, MUSE topline and POLY baseline) and we were ranked 2nd in the shared task
Keywords: Summarization, Neural networks, Reinforcement learning, sequence to sequence learning; actor-critic methods; policy gradients.