Emerson – A Conversational AI

The project intends to create a conversational AI, Emerson, who will be able to detect users’ intent, drive further interactions on different topics, including providing knowledge-based information as requested, discussing opinions on popular issues, and providing suggestions on issues related to with several subcomponents. We envision Emerson as an aggregation of domain-specific information agents, together with a dialogue management component, and an auxiliary reinforcement learning system to conduct conversations with users. The architecture of Emerson and an example of conversation flow are illustrated in the following figures.alexa-prize-architecture

 

Conversation-Example

Conversation-Example

Group Photo

Team Members

Tomasz Jurczyk

Payam Karisani

Garry Lai

Bonggun Shin

upload

Zihao Wang(Project Lead)

Ethan Zhou

Advisor

Eugene Agichtein