Phase Coherence in Conceptual Spaces for Conversational Agents

This chapter attempts to enhance the traditional chatbots with associative/intuitive capabilities. According to these considerations, it tries to create a conversational agent model that takes into consideration, aside from the traditional rule - based dialogue mechanism, also some sort of intuitive reasoning ability. The aim is in attempting to overcome the rigid pattern - matching rules, proposing a "phase coherence" paradigm into a semantic space. With this locution the chapter intend that the vectors representing the elements of the dialogue are coherent with the context. The chapter trust that this intuitive - associative capability can be obtained using the LSA methodology. The representation of information in a LSA - based semantic, "conceptual," manifold and the resulting subsymbolic geometric representation of the chatbot knowledge can contribute to better design a humanlike conversational interface provided with both intuitive - associative capabilities and a rulebased dialogue skill.