Across the board: Human-human game learning dialogues

Abstract:

Learning novel tasks through dialogue and interaction is a common human activity. The field of Interactive Task Learning looks to give this capability to AI systems. However, the dialogue strategies used in those systems are not based on observed human strategies. This paper presents a data collection of instructive dialogue in the form of filmed interactions where a teacher teaches a board game to a student.

Research areas:

• Dialogue
• DivCon

Year:

2025

Type of Publication:

In Proceedings

Book title:

Proceedings of the 29th Workshop on the Semantics and Pragmatics of Dialogue - Poster Abstract.