PoinT-5: Pointer Network and T-5 based Financial NarrativeSummarisation
This addresses the need for automated summarization of long financial documents for stakeholders, but it is incremental as it builds on existing techniques like Pointer Networks and T-5.
The paper tackles the problem of summarizing lengthy financial reports by proposing PoinT-5, a method combining Pointer Networks and T-5 to extract and paraphrase sentences, achieving the highest precision scores in ROUGE metrics and crossing the MUSE baseline in ROUGE-LCS.
Companies provide annual reports to their shareholders at the end of the financial year that describes their operations and financial conditions. The average length of these reports is 80, and it may extend up to 250 pages long. In this paper, we propose our methodology PoinT-5 (the combination of Pointer Network and T-5 (Test-to-text transfer Transformer) algorithms) that we used in the Financial Narrative Summarisation (FNS) 2020 task. The proposed method uses pointer networks to extract important narrative sentences from the report, and then T-5 is used to paraphrase extracted sentences into a concise yet informative sentence. We evaluate our method using ROUGE-N (1,2), L, and SU4. The proposed method achieves the highest precision scores in all the metrics and highest F1 scores in ROUGE1, and LCS and the only solution to cross the MUSE solution baseline in ROUGE-LCS metrics.