A Generative Approach for Financial Causality Extraction

TL;DR

This paper introduces a generative encoder-decoder approach with pointer networks for extracting multiple and overlapping causality relations from financial texts, outperforming previous sequence labeling methods.

Contribution

It proposes a novel generative framework for financial causality extraction that effectively handles multiple and overlapping causalities, improving over prior sequence labeling techniques.

Findings

Achieves competitive performance on the FinCausal dataset.

Effectively extracts multiple causalities from financial texts.

Handles overlapping causality spans better than previous methods.

Abstract

Causality represents the foremost relation between events in financial documents such as financial news articles, financial reports. Each financial causality contains a cause span and an effect span. Previous works proposed sequence labeling approaches to solve this task. But sequence labeling models find it difficult to extract multiple causalities and overlapping causalities from the text segments. In this paper, we explore a generative approach for causality extraction using the encoder-decoder framework and pointer networks. We use a causality dataset from the financial domain, \textit{FinCausal}, for our experiments and our proposed framework achieves very competitive performance on this dataset.