학술논문
Neural Generation Meets Real People: Building a Social, Informative Open-Domain Dialogue Agent
Document Type
Working Paper
Author
Chi, Ethan A.; Paranjape, Ashwin; See, Abigail; Chiam, Caleb; Chang, Trenton; Kenealy, Kathleen; Lim, Swee Kiat; Hardy, Amelia; Rastogi, Chetanya; Li, Haojun; Iyabor, Alexander; He, Yutong; Sowrirajan, Hari; Qi, Peng; Sadagopan, Kaushik Ram; Phu, Nguyet Minh; Soylu, Dilara; Tang, Jillian; Narayan, Avanika; Campagna, Giovanni; Manning, Christopher D.
Source
Subject
Language
Abstract
We present Chirpy Cardinal, an open-domain social chatbot. Aiming to be both informative and conversational, our bot chats with users in an authentic, emotionally intelligent way. By integrating controlled neural generation with scaffolded, hand-written dialogue, we let both the user and bot take turns driving the conversation, producing an engaging and socially fluent experience. Deployed in the fourth iteration of the Alexa Prize Socialbot Grand Challenge, Chirpy Cardinal handled thousands of conversations per day, placing second out of nine bots with an average user rating of 3.58/5.
Comment: SIGDIAL '22
Comment: SIGDIAL '22