Training an adaptive dialogue policy for interactive learning of visually grounded word meanings

Yu, Yanchao; Eshghi, Arash; Lemon, Oliver

Abstract

We present a multi-modal dialogue system for interactive learning of perceptually grounded word meanings from a human tutor. The system integrates an incremental, semantic parsing/generation framework - Dynamic Syntax and Type Theory with Records (DS-TTR) - with a set of visual classifiers that are learned throughout the interaction and which ground the meaning representations that it produces. We use this system in interaction with a simulated human tutor to study the effects of different dialogue policies and capabilities on accuracy of learned meanings, learning rates, and efforts/costs to the tutor. We show that the overall performance of the learning agent is affected by (1) who takes initiative in the dialogues; (2) the ability to express/use their confidence level about visual attributes; and (3) the ability to process elliptical and incrementally constructed dialogue turns. Ultimately, we train an adaptive dialogue policy which optimises the trade-off between classifier accuracy and tutoring costs.

Citation

Yu, Y., Eshghi, A., & Lemon, O. (2016, September). Training an adaptive dialogue policy for interactive learning of visually grounded word meanings. Presented at 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue, Los Angeles, US

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue
• Start Date: Sep 13, 2016
• End Date: Sep 15, 2016
• Publication Date: 2016
• Deposit Date: Jun 28, 2023
• Publicly Available Date: Jun 28, 2023
• Publisher: Association for Computational Linguistics (ACL)
• Pages: 339-349
• Book Title: Proceedings of the 17th Annual Meeting of the Special Interest Group on Discourse and Dialogue
• DOI: https://doi.org/10.18653/v1/w16-3643
• Publisher URL: https://doi.org/10.18653/v1/w16-3643