Language-Action Tools for Cognitive Artificial Agents: Integrating Vision, Action and Language
Papers from the 2011 AAAI Workshop
Katerina Pastra, Yiannis Aloimonos, Workshop Cochairs
Endowing artificial agents with language and action abilities has been a quest in many AI subfields. A number of AI applications require coupling of language and motoric or visual action (and objects), ranging from language-based human-robot interaction to event recognition. Recent years have witnessed great advances in different disciplines that provide the theoretical and technological framework for an interdisciplinary approach to language-action integration. Neuroscience research provides more and more evidence on a common neural basis for language and action, both in perception and in production. A growing body of experimental cognitive science findings sheds light on the close interaction and reciprocal influence of language and action in a number of tasks, such as categorization and learning. On their part, technological advances in multisensory human behaviour measurement have enabled the development of recognitive and generative algorithms for the analysis of sensorimotor representations, ones that are analogical to language analysis and generation models. How could language-action integration benefit from all such developments? If it does, will this lead beyond the current state of the art in real-life AI applications that require generalization and optimality in language-action integration?
The goal of this workshop is to bring together an interdisciplinary group of computational linguists, computer vision researchers, roboticists and neuroscientists that will address the issue of developing biologically-inspired language and action technology for artificial agents. The focus is both on (1) how individual technologies can benefit from interdisciplinary research for going beyond the state of art in language-action integration tasks, and (2) how language processing, visual processing and/or motor control algorithms can be integrated in artificial agents allowing for behaviour generalization and optimization.