Vision-and-Language or Vision-for-Language? On Cross-Modal Influence in Multimodal Transformers

Stella Frank, Emanuele Bugliarello, Desmond Elliott


Abstract
Pretrained vision-and-language BERTs aim to learn representations that combine information from both modalities. We propose a diagnostic method based on cross-modal input ablation to assess the extent to which these models actually integrate cross-modal information. This method involves ablating inputs from one modality, either entirely or selectively based on cross-modal grounding alignments, and evaluating the model prediction performance on the other modality. Model performance is measured by modality-specific tasks that mirror the model pretraining objectives (e.g. masked language modelling for text). Models that have learned to construct cross-modal representations using both modalities are expected to perform worse when inputs are missing from a modality. We find that recently proposed models have much greater relative difficulty predicting text when visual information is ablated, compared to predicting visual object categories when text is ablated, indicating that these models are not symmetrically cross-modal.
Anthology ID:
2021.emnlp-main.775
Volume:
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2021
Address:
Online and Punta Cana, Dominican Republic
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
9847–9857
Language:
URL:
https://aclanthology.org/2021.emnlp-main.775
DOI:
10.18653/v1/2021.emnlp-main.775
Bibkey:
Cite (ACL):
Stella Frank, Emanuele Bugliarello, and Desmond Elliott. 2021. Vision-and-Language or Vision-for-Language? On Cross-Modal Influence in Multimodal Transformers. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 9847–9857, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Vision-and-Language or Vision-for-Language? On Cross-Modal Influence in Multimodal Transformers (Frank et al., EMNLP 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.emnlp-main.775.pdf
Video:
 https://aclanthology.org/2021.emnlp-main.775.mp4
Code
 e-bug/volta +  additional community code
Data
Conceptual CaptionsFlickr30k