Am I capable yet? ...where I argue we need better evaluation and better representations to improve cognitive architectures for HRI
Séverin Lemaignan · Infoscience (Ecole Polytechnique Fédérale de Lausanne) · 2014
I suggest in [1] that the cognitive foundations that are required to further advance research on cognitive architec-tures for human-robot interaction must answer these three questions: 1. How can I represent my environment and my beliefs in such a way that perception modalities as diverse as my own sensors, verbal descriptions of distant ob-jects or events, or symbolic common-sense knowledge extracted from online bases could be mixed? 2. How do I identify the current situation under a com-plex combination of diverse, changing circumstances? How do I project myself in other situations to predict, remember, explain, learn? 3. How do I access and represent mental models of the agents I interact with?