Beyond Image-Text Matching: Verb Understanding in Multimodal Transformers Using Guided Masking
This provides a more accurate probing method for evaluating multimodal AI models, addressing a specific limitation in understanding verbs, though it is incremental in improving evaluation techniques.
The paper tackles the problem of assessing verb understanding in multimodal transformers, showing that guided masking reveals these models can predict verbs with high accuracy, contrasting with previous image-text matching methods that often fail.
The dominant probing approaches rely on the zero-shot performance of image-text matching tasks to gain a finer-grained understanding of the representations learned by recent multimodal image-language transformer models. The evaluation is carried out on carefully curated datasets focusing on counting, relations, attributes, and others. This work introduces an alternative probing strategy called guided masking. The proposed approach ablates different modalities using masking and assesses the model's ability to predict the masked word with high accuracy. We focus on studying multimodal models that consider regions of interest (ROI) features obtained by object detectors as input tokens. We probe the understanding of verbs using guided masking on ViLBERT, LXMERT, UNITER, and VisualBERT and show that these models can predict the correct verb with high accuracy. This contrasts with previous conclusions drawn from image-text matching probing techniques that frequently fail in situations requiring verb understanding. The code for all experiments will be publicly available https://github.com/ivana-13/guided_masking.