AIDA: AI for Financial Document Analysis

AIDA (Intelligent Analysis of Financial Asset Documents) is a research project focused on developing NLP and machine learning methods for the automatic analysis of financial documents. The project targets the extraction, interpretation, and summarization of information from complex financial texts such as earnings reports, regulatory filings, prospectuses, and market disclosures.

By leveraging recent advances in large language models and knowledge-enriched representations, AIDA aims to build intelligent tools that support financial analysts, investors, and regulators in navigating the growing volume of textual financial data — with a particular focus on the Brazilian financial market and Portuguese-language documents.

Objectives

• Develop models for generation, information extraction, classification, and summarization of financial texts
• Build and release annotated datasets and benchmarks for Portuguese-language financial NLP
• Investigate the use of domain-adapted language models for financial document understanding
• Support decision-making in financial markets through explainable and reliable AI tools

Funding & Support

This project is financed by CNPq in partnership with IAIA (Institute for Artificial Intelligence), covering the period from June 2024 to June 2026.