A hybrid approach to learn, retrieve and reuse qualitative cases

Autor(es)

Homem, Thiago P. D. ; Perico, Danilo H. ; Santos, Paulo E. ; Costa, Anna H. R. ; Bianchi, Reinaldo A. C. ; de Mantaras, Ramon Lopez ; Todt, E ; Tonidandel, F

• Tipo de documento: Artigo Científico
• Fonte: 2017 LATIN AMERICAN ROBOTICS SYMPOSIUM (LARS) AND 2017 BRAZILIAN SYMPOSIUM ON ROBOTICS (SBR); v. N/A, p. 6-pg., 2017-01-01.

Resumo

The application of Artificial Intelligence methods is becoming indispensable in several domains, for instance in credit card fraud detection, voice recognition, autonomous cars and robotics. However, some methods fail in performances or solving some problems, and hybrid approaches can outperform the results when compared to traditional ones. In this paper we present a hybrid approach, named qualitative case-based reasoning and learning (QCBRL), that integrates three well-known AI methods: Qualitative Spatial Reasoning, Case-Based Reasoning and Reinforcement Learning. QCBRL system was designed to allow an agent to learn, retrieve and reuse qualitative cases in the robot soccer domain. We applied our method in the Half-Field Offense and we have obtained promising results. (AU)

• Processo FAPESP: 16/21047-3 - ALIS: Aprendizado Autônomo em Sistemas Inteligentes
• Beneficiário: Anna Helena Reali Costa
• Modalidade de apoio: Auxílio à Pesquisa - Regular
• Processo FAPESP: 16/18792-9 - Descrição, representação e solução de jogos espaciais
• Beneficiário: Paulo Eduardo Santos
• Modalidade de apoio: Auxílio à Pesquisa - Parceria para Inovação Tecnológica - PITE