In English

ChatUp; En Chatbot med kontinuerligt lyssnande samt snabb respons.

Johan Martinson ; Daniel Rydén
Göteborg : Chalmers tekniska högskola, 2018. 44 s.
[Examensarbete på grundnivå]

Interest in IT and especially artificial intelligence increases in Swedish companies by the day. Therefore a platform in the form of a robot to accommodate experiments in machine learning and machine intelligence has been developed. The idea is that this basic product will capture peoples’ interests during fairs and events. The developed software module ChatUp has been integrated into a Nao-robot, which is a small humalike robot, that takes in user requests, interprets them via a NLP and then selects an appropriate service to finally respond. Since the idea is that the robot should behave as humanlike as possible, the time between request and response should be shortened to a minimum. The project has therefore investigated sound segmentation algorithms, to remove silence and non-speech in audio this with the intention of minimizing the size of its data packet transmitted over the network, thus shortening the response time. The investigation has resulted in the fact that manipulation of live-streaming audio is not suitable while live streaming analysis can shorten response time noticeably. The results will be used as a basis for further development of the robot and the ChatUp module.

Nyckelord: NLP, audio segmentation algorithm, response time, chatbot, artificial intelligence, zero crossing rate, short term energy, live-streaming



Publikationen registrerades 2018-07-02. Den ändrades senast 2018-07-02

CPL ID: 255390

Detta är en tjänst från Chalmers bibliotek