Leverage Points in Modality Shifts: Comparing Language-only and Multimodal Word Representations
This is an incremental study that provides insights into how visual modality influences language representations for researchers in multimodal AI and computational linguistics.
The paper compared word embeddings from vision-and-language models and text-only models to identify semantic properties most affected by visual grounding, finding that effects correlate strongly with concreteness and also appear in specific semantic classes and sentiment-related valence.
Multimodal embeddings aim to enrich the semantic information in neural representations of language compared to text-only models. While different embeddings exhibit different applicability and performance on downstream tasks, little is known about the systematic representation differences attributed to the visual modality. Our paper compares word embeddings from three vision-and-language models (CLIP, OpenCLIP and Multilingual CLIP) and three text-only models, with static (FastText) as well as contextual representations (multilingual BERT; XLM-RoBERTa). This is the first large-scale study of the effect of visual grounding on language representations, including 46 semantic parameters. We identify meaning properties and relations that characterize words whose embeddings are most affected by the inclusion of visual modality in the training data; that is, points where visual grounding turns out most important. We find that the effect of visual modality correlates most with denotational semantic properties related to concreteness, but is also detected for several specific semantic classes, as well as for valence, a sentiment-related connotational property of linguistic expressions.