Dialogue Systems embrace the ultimate goal of human-robot interaction, in which computational systems communicate with their human interlocutors in the same way humans communicate among them. Although significant progress has been achieved during the last few years and some pioneering commercial systems are finding already their way in the market, current state-of-the-art dialogue systems are very limited in their approach to the human communication phenomenon. Some of these limitations include:

• the lack of ability to properly model world knowledge for reasoning purposes,
• the still low reliability of speech recognition and the limited capacity of dialogue systems to cope with recognition errors in a logical manner,
• the complex and multimodal nature of the pragmatic phenomena and the socio-cultural aspects that affect such interactions, and
• the difficulty of evaluating dialogue quality and the subsequent development of reliable strategies for automatic learning and adaptation.

The main objective of this NII-Shonan meeting is to discuss about the most relevant and promising future directions of research in dialogue systems. The discussion is to be centered on how these research directions address the different problems and limitations of current dialogue systems, as well as how they provide the basis for the next generation of intelligent artificial agents.

The assistants will be requested to present work in progress as well as to propose new collaborative efforts for tackling with the main limitations of current dialogue systems mentioned above. As a result of the meeting, a research agenda for main directions and international collaboration for the next three to five years will be defined.

Joint organization of future workshops and shared tasks by the participants in the meeting will be defined with the objective of pushing the state-of-the-art in dialogue systems to a more comprehensive effort for developing a new generation of intelligent virtual assistants.