Spatial reasoning from the point of view of the first robot" seminar

The ability to perceive space and reason about spatio-temporal relationships is effortless for humans, but it has proven to be a challenge for computer systems that struggle to process the various nuances of our conceptualization of the world.

This talk presents two results on the development of qualitative spatial reasoning tools applied to multi-robot systems that aim to bridge the gap between how humans and machines interpret and act on the external world.

The first is a new algorithm for Qualitative Case-Based Reasoning and Learning (QCBRL), which is a case-based reasoning system that uses qualitative spatial representations to retrieve and reuse cases by means of relationships between objects in the environment. Combined with reinforcement learning, QCBRL enables the agent to learn new qualitative cases at runtime, without going through a pre-processing stage. Experimental evaluation of QCBRL was carried out in a simulated robot-soccer environment and in a real humanoid robot. The results show that QCBRL outperforms traditional RL methods and state-of-the-art CBR systems.

The second result is an algorithm for combining information obtained from multiple (distinct and egocentric) viewpoints to infer the position, route and actions needed to guide a sensory-deprived agent to a given destination. Information from multiple observers has been merged in terms of a set of qualitative directions that can be easily interpreted by a human agent.

Keywords: Qualitative spatial reasoning, case-based reasoning, multi-robot systems.

Paulo Santos received his PhD in Artificial Intelligence from Imperial College London, UK, where he worked on the development of spatial reasoning systems for mobile robots. He was a research assistant at the School of Computing, University of Leeds, UK, where he worked on an EU-funded project for the development of cognitive vision systems. Mr. Santos led an AI and robotics research group in Sao Paulo, Brazil (2005-2019), conducting a number of research projects of industrial interest. During this period, Mr. Santos was also a visiting researcher at the following world-renowned institutions: University of Leeds, UK (2007, 2010); Ryerson University, Canada (2010), University of Bremen, Germany (2012); University of La Coruña, Spain (2014). Dr. Santos currently holds the position of Associate Professor of AI and Robotics at Flinders University, Adelaide, Australia, since 2020, developing research on the integration of automated reasoning with deep learning models towards explainable AI systems.