Interactively learning visually grounded word meanings from a human tutor

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 effect 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 as well as incrementally constructed dialogue turns.

Citation

Yu, Y., Eshghi, A., & Lemon, O. (2016, August). Interactively learning visually grounded word meanings from a human tutor. Presented at 5th Workshop on Vision and Language, Berlin, Germany

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 5th Workshop on Vision and Language
• Start Date: Aug 12, 2016
• Publication Date: 2016
• Deposit Date: Jun 28, 2023
• Publicly Available Date: Jun 28, 2023
• Publisher: Association for Computational Linguistics (ACL)
• Pages: 48-53
• Book Title: Proceedings of the 5th Workshop on Vision and Language
• Publisher URL: https://aclanthology.org/W16-3206/