Information Extraction through AI techniques: The KIDs use case at CONSOB

TL;DR

This paper discusses initial efforts to automate information extraction from financial documents using rule-based and machine learning methods in a collaboration between CONSOB and Sapienza University.

Contribution

It introduces a hybrid approach combining rule-based and machine learning techniques for extracting financial information from documents.

Findings

Initial results demonstrate feasibility of automated extraction

Hybrid methods outperform purely rule-based approaches

Framework applicable to financial document analysis

Abstract

In this paper we report on the initial activities carried out within a collaboration between Consob and Sapienza University. We focus on Information Extraction from documents describing financial instruments. We discuss how we automate this task, via both rule-based and machine learning-based methods and provide our first results.