Hello, We are pleased to propose a 5 month second year master internship at the LIRMM laboratory in Montpellier (south of France) investigating questions of embodiment in human robot interaction (see description below). The internship will be jointly supervised by Madalina Croitoru (Computer Science) and Ganesh Gowrishankar (Robotics and Neuroscience).
The start date is February 2023 but please note that non French candidates must complete some administrative procedures lasting up to two months before starting the internship. If you are interested please get in touch by email by the latest the end of December, Thanks, Madalina Since the 50’s Artificial Intelligence has captured imaginations with the dream of developing software displaying human-like behaviour. First formalised as the Turing Test, such behaviour was first and foremost conversational. A human being carried out a dialogue (without being able to physically see its partner) and correctly guessed whether the partner is natural or artificial. Since the breakthrough of pioneering chatbots passing the Turing test, natural language models have greatly evolved today to the point that the question of their very sentience becomes a societal concern. At the same time, large companies spend the equivalent of a small country's budget investing in the development of XR (eXtended Reality - capturing both virtual and augmented reality). While a lot of research has been dedicated to the technical aspect of physical interactions with such software, less research has focused on the cognitive aspects. In this work we restrict ourselves to the realm of dialogue as means of cognitive interaction. We follow classic argumentation based distinctions and restrict our focus on three types of dialogues: persuasion, negotiation and inquiry. Contrary to existing work we do not focus on the technical challenges of making the machine “understand” the dialogue. We are interested in the question of how the embodiment of the machine impacts the humanly perceived quality of the interaction. Much research has been dedicated to how the physical embodiment of robots affects the quality of its interaction with the human counterpart. In order to restrict the relevant literature we focus on cognitive interactions. Even so, questions of visual perception, physical ability, text to speech software, emotional abilities etc. of the artificial agent play a major proven role in the quality of the perceived interaction of the human. Several other factors play in as well such as the human’s age, cognitive abilities, gender and cultural background Against this background the research question this internship will consist of investigating whether the physical embodiment of the artificial agent impacts the perceived quality of the cognitive interaction with the human. We will restrict ourselves to the case of two humanoid like robots: Pepper and QTRobot and study three kinds of dialogues (persuasion, negotiation and information seeking) of humans either with the actual robot or with a VR depiction of the robot. The internship will consist of several deliverables: (1) a state of the art of the various factors affecting the human perceived quality of cognitive interaction in the context of dialogues; (2) a first study of physical embodiment in the context of predefined dialogues and (3) a study of physical embodiment impact in the context of “free-range” dialogues. -- Madalina Croitoru, Professor BOREAL INRIA, LIRMM Deputy director of the Computer Science Department, Faculty of Science, University of Montpellier
_______________________________________________ uai mailing list uai@engr.orst.edu https://it.engineering.oregonstate.edu/mailman/listinfo/uai
