Do You See What I Mean? Visual Resolution of Linguistic Ambiguities

TL;DR

This paper introduces a new multimodal task for grounded language understanding, where visual scenes are used to disambiguate ambiguous sentences, supported by a novel corpus and an extended vision model.

Contribution

It presents a new dataset of ambiguous sentences with visual interpretations and extends a vision model to disambiguate sentences across various ambiguity types.

Findings

Model successfully disambiguates sentences using visual context

New corpus captures diverse syntactic, semantic, and discourse ambiguities

Approach unifies disambiguation across different ambiguity types

Abstract

Understanding language goes hand in hand with the ability to integrate complex contextual information obtained via perception. In this work, we present a novel task for grounded language understanding: disambiguating a sentence given a visual scene which depicts one of the possible interpretations of that sentence. To this end, we introduce a new multimodal corpus containing ambiguous sentences, representing a wide range of syntactic, semantic and discourse ambiguities, coupled with videos that visualize the different interpretations for each sentence. We address this task by extending a vision model which determines if a sentence is depicted by a video. We demonstrate how such a model can be adjusted to recognize different interpretations of the same underlying sentence, allowing to disambiguate sentences in a unified fashion across the different ambiguity types.