Development of an attention module for CST - The Cognitive Systems Toolkit

• Grant number: 18/06714-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2018
• End date: September 30, 2018
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Agreement: IBM Brasil
• Principal Investigator: Ricardo Ribeiro Gudwin
• Grantee: Leandro Nascimento Gonçalves de Araújo
• Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Company: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação (FEEC)
• Associated research grant: 16/18819-4 - Bootstrap conditions for interaction-based multimodal learning in cognitive robotics, AP.PITE

Abstract

The objective of this scientific initiation is the study and development of attentional modules to be incorporated to CST [1], in order to provide support for the project "Bootstrap Conditions for Interaction-Based Multimodal Learning in Cognitive Robots". In this project, the aim is at looking for the minimal conditions for cognition - what we are calling here "bootstrap conditions", to allow a cognitive system, like a humanoid robot, to interact with the world and learn from this interaction. A member of our research group, Prof. Esther Colombini, has developed in her PdD Thesis [2] a sophisticated general attention model [3] to be incorporated in different cognitive architectures. During her PhD, she created a C++ platform with an implementation of such model, which was used for her tests and simulations within the thesis. Our research group is fostering the development of CST, the "Cognitive Systems Toolkit", a general toolkit for the construction of cognitive architectures, using and integrating many different technologies. Our intention with the current scientific initiation is to port Prof. Colombini's developments to CST, with the required adaptations and transpositions necessary to incorporate her models with the CST Core philosophy. The original implementation was made in C++, and biased to the experiments she developed for her thesis. Our aim here is to turn this reference code into a flexible and reusable library, consolidated in a set of modules within CST. Once ported to CST, this attention module will be used by others in the experiments involved in the project. (AU)