Using Deep Learning and an External Knowledge Base to Develop Human-Robot Dialogues

To achieve human-robot communication in a more natural way, in this work we develop a human-machine dialogue system to provide domain-specific knowledge services. The conversational modeling is regarded as learning a model to conduct mapping between human utterances and machine responses, and a deep learning neural model is adopted to perform answer selection. Our system also includes an external knowledge resource to further enrich knowledge for searching answers. Moreover, a reasoning procedure is constructed to look for semantically similar questions from the built-in knowledge repository, and then to retrieve their answers. Extensive sets of experiments are conducted and the results show the promises and potentials of the presented system.