Investigation of Question Answering Algorithms in the BigQA Software Reference Architecture

• Grant number: 23/08293-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: August 01, 2023
• End date: July 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Cristina Dutra de Aguiar
• Grantee: Pedro Calciolari Jardim
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

Aiming to investigate the challenge of building a Big Data Question Answering), the research group conducted by this project supervisor is developing the BigQA software reference architecture. Among its functionalities, BigQA encompasses a module that supports Question Answering algorithms. These algorithms are composed of two steps. The first step, named document retriever, receives questions written in natural language as input and searches the documents to find helpful information in response to the questions. The second step, named document reader, analyzes the obtained information and generates an answer as output. This project aims to investigate employing different QA algorithms to implement the document reader of the BigQA architecture. To comply with this goal, we aim to indicate the most appropriate QA algorithm according to the characteristics of the questions and the results returned by the document retriever. We also intend to generate a dataset that allows the comparison of the QA algorithms considering big data scenarios.