Reference-Centric Models for Grounded Collaborative Dialogue

Citations Over TimeTop 14% of 2021 papers

Abstract

We present a grounded neural dialogue model that successfully collaborates with people in a partially-observable reference game. We focus on a setting where two agents each observe an overlapping part of a world context and need to identify and agree on some object they share. Therefore, the agents should pool their information and communicate pragmatically to solve the task. Our dialogue agent accurately grounds referents from the partner's utterances using a structured reference resolver, conditions on these referents using a recurrent memory, and uses a pragmatic generation procedure to ensure the partner can resolve the references the agent produces. We evaluate on the OneCommon spatial grounding dialogue task (Udagawa and Aizawa, 2019), involving a number of dots arranged on a board with continuously varying positions, sizes, and shades. Our agent substantially outperforms the previous state of the art for the task, obtaining a 20% relative improvement in successful task completion in self-play evaluations and a 50% relative improvement in success in human evaluations.

Related Papers

• → Reconsideration of the simulated work task situation(2010)38 cited
• → THE RELATION BETWEEN CHOOSING AND WORKING PREVOCATIONAL TASKS IN TWO SEVERELY RETARDED YOUNG ADULTS(1980)76 cited
• → Comparing Two Objects for Size after a Comparison of One of the Objects with other Objects(1980)1 cited
• Task Patterns for Human-Robot Interaction(2002)
• → MG313 Daily life support robot which picks up an object indicated by a human(2007)