Visually Grounded Speech Models Have a Mutual Exclusivity Bias

Leanne Nortje, Dan Oneaţă, Yevgen Matusevych, Herman Kamper


Abstract
When children learn new words, they employ constraints such as the mutual exclusivity (ME) bias: A novel word is mapped to a novel object rather than a familiar one. This bias has been studied computationally, but only in models that use discrete word representations as input, ignoring the high variability of spoken words. We investigate the ME bias in the context of visually grounded speech models that learn from natural images and continuous speech audio. Concretely, we train a model on familiar words and test its ME bias by asking it to select between a novel and a familiar object when queried with a novel word. To simulate prior acoustic and visual knowledge, we experiment with several initialization strategies using pretrained speech and vision networks. Our findings reveal the ME bias across the different initialization approaches, with a stronger bias in models with more prior (in particular, visual) knowledge. Additional tests confirm the robustness of our results, even when different loss functions are considered. Based on detailed analyses to piece out the model’s representation space, we attribute the ME bias to how familiar and novel classes are distinctly separated in the resulting space.
Anthology ID:
2024.tacl-1.42
Volume:
Transactions of the Association for Computational Linguistics, Volume 12
Month:
Year:
2024
Address:
Cambridge, MA
Venue:
TACL
SIG:
Publisher:
MIT Press
Note:
Pages:
755–770
Language:
URL:
https://aclanthology.org/2024.tacl-1.42
DOI:
10.1162/tacl_a_00672
Bibkey:
Cite (ACL):
Leanne Nortje, Dan Oneaţă, Yevgen Matusevych, and Herman Kamper. 2024. Visually Grounded Speech Models Have a Mutual Exclusivity Bias. Transactions of the Association for Computational Linguistics, 12:755–770.
Cite (Informal):
Visually Grounded Speech Models Have a Mutual Exclusivity Bias (Nortje et al., TACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.tacl-1.42.pdf