Topical-Chat: Towards Knowledge-Grounded Open-Domain Conversations
This addresses the problem of limited datasets for knowledge-grounded conversational AI, though it is incremental as it builds on existing work by expanding topical coverage.
The authors tackled the challenge of building socialbots for deep open-domain conversations by introducing Topical-Chat, a human-human dataset with knowledge spanning 8 topics and no explicit roles, and they trained state-of-the-art models on it for benchmarking.
Building socialbots that can have deep, engaging open-domain conversations with humans is one of the grand challenges of artificial intelligence (AI). To this end, bots need to be able to leverage world knowledge spanning several domains effectively when conversing with humans who have their own world knowledge. Existing knowledge-grounded conversation datasets are primarily stylized with explicit roles for conversation partners. These datasets also do not explore depth or breadth of topical coverage with transitions in conversations. We introduce Topical-Chat, a knowledge-grounded human-human conversation dataset where the underlying knowledge spans 8 broad topics and conversation partners don't have explicitly defined roles, to help further research in open-domain conversational AI. We also train several state-of-the-art encoder-decoder conversational models on Topical-Chat and perform automated and human evaluation for benchmarking.