Language Bias-Driven Self-Knowledge Distillation with Generalization Uncertainty for Reducing Language Bias in Visual Question Answering

To answer questions, visual question answering systems (VQA) rely on language bias but ignore the information of the images, which has negative information on its generalization. The mainstream debiased methods focus on removing language prior to inferring. However, the image samples are distributed...

Descripció completa

Dades bibliogràfiques

Autors principals:	Desen Yuan, Lei Wang, Qingbo Wu, Fanman Meng, King Ngi Ngan, Linfeng Xu
Format:	Artigo
Idioma:	Inglês
Publicat:	MDPI AG 2022-07-01
Col·lecció:	Applied Sciences
Matèries:	visual question answering self-knowledge distillation language bias generalized uncertainty
Accés en línia:	https://www.mdpi.com/2076-3417/12/15/7588
Etiquetes:	Afegir etiqueta Sense etiquetes, Sigues el primer a etiquetar aquest registre!