A motivation-based incremental learning framework for robotics

• Grant number: 21/07050-0
• Support Opportunities: Scholarships in Brazil - Doctorate
• Effective date (Start): April 01, 2022
• Effective date (End): June 04, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Esther Luna Colombini
• Grantee: Letícia Mara Berto
• Host Institution: Instituto de Computação (IC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 13/07559-3 - BRAINN - The Brazilian Institute of Neuroscience and Neurotechnology, AP.CEPID

Abstract

Recently, the use of robots in daily activities has become common. As a result, the environments are more complex, demanding the robots develop the learning ability through interaction with the environment and other agents. This capacity allows robots to execute activities that have not been pre-programmed. For this, robots must have cognition, which can convert information from different sources into knowledge, a process done through cognitive functions. Each cognitive function has its specificity, but they need to act together for development. However, it is reasonable to assume that some previous conditions must be achieved before complex abilities can emerge. In recent approaches, some researchers have hypothesized that some minimal conditions exist for cognition to evolve.Another critical point is that emotion and motivation are significant impacts on decision-making and human behavior. Therefore they are essential functions to be developed in artificial agents. Our research is in the context of developmental and cognitive robotics, which aim to build robots based on human development using cognitive architectures that represent comprehensive computer models providing theoretical frameworks to work with cognitive processes in the search for complex behavior. Therefore, continuing the study carried out in my master's degree, we will design a cognitive agent with emotional and motivational systems, modeling the minimum necessary cognitive processes and their activation dynamics. To validate our agent's learning, we will run a series of incremental experiments that the agent can carry out individually or with social interaction. (AU)