World-to-Words: Grounded Open Vocabulary Acquisition through Fast Mapping in Vision-Language Models

TL;DR

This paper introduces OctoBERT, a visually-grounded language model that learns grounded word meanings rapidly and robustly, demonstrating improved open vocabulary acquisition through grounding and fast mapping in vision-language models.

Contribution

It proposes OctoBERT, a novel pre-trained vision-language model that enhances grounded word learning and bootstrapping of unseen words in open-world settings.

Findings

OctoBERT outperforms baselines in grounded word learning.

Grounding during pre-training accelerates unseen word acquisition.

OctoBERT demonstrates more coherent and rapid open vocabulary learning.

Abstract

The ability to connect language units to their referents in the physical world, referred to as grounding, is crucial to learning and understanding grounded meanings of words. While humans demonstrate fast mapping in new word learning, it remains unclear whether modern vision-language models can truly represent language with their grounded meanings and how grounding may further bootstrap new word learning. To this end, we introduce Grounded Open Vocabulary Acquisition (GOVA) to examine grounding and bootstrapping in open-world language learning. As an initial attempt, we propose object-oriented BERT (OctoBERT), a novel visually-grounded language model by pre-training on image-text pairs highlighting grounding as an objective. Through extensive experiments and analysis, we demonstrate that OctoBERT is a more coherent and fast grounded word learner, and that the grounding ability acquired during pre-training helps the model to learn unseen words more rapidly and robustly. Our code is available at https://github.com/sled-group/world-to-words