Visually Grounded Speech Models for Low-resource Languages and Cognitive Modelling
This work addresses challenges in low-resource language processing and cognitive science, though it appears incremental by building on existing visually grounded speech models.
The dissertation tackles the problem of learning from unlabelled speech paired with images for low-resource languages and cognitive modelling, introducing a visually prompted keyword localisation task and demonstrating effectiveness in few-shot learning for languages like Yoruba, with results showing a monolingual model exhibits mutual exclusivity bias while multilingualism does not affect it similarly to children.
This dissertation examines visually grounded speech (VGS) models that learn from unlabelled speech paired with images. It focuses on applications for low-resource languages and understanding human language acquisition. We introduce a task called visually prompted keyword localisation to detect and localise keywords in speech using images. We demonstrate the effectiveness of VGS models in few-shot learning scenarios for low-resource languages like Yoruba. Additionally, we examine the mutual exclusivity bias in VGS models. Our monolingual VGS model exhibits this bias, but we found that multilingualism does not affect the bias in this VGS model similarly to what is observed in children.