Work

HQD-EM: Robust VQA Through Hierarchical Question Decomposition Bias Module and Ensemble Adaptive Angular Margin Loss

Year: 2025

Type: article

Abstract: Recent studies in Visual Question Answering (VQA) have revealed that models often rely heavily on language priors rather than vision-language understanding, leading to poor generalization under distribution shifts. To address this challenge, we propose HQD-EM, a unified debiasing framework that combines Hierarchical Question Decomposition module (HQD) with an Ensemble adaptive angular Margin loss (EM). HQD systematically decomposes questions into multi-granular representations to capture layered language biases, while EM leverages bias confidence to modulate per-sample decision margins dynamically. Our method integrates generative bias modeling with adaptive margin learning in an end-to-end trainable architecture. Experiments on VQA benchmarks demonstrate that HQD-EM significantly outperforms prior work and achieves state-of-the-art results on VQA-CP2 and VQA-CP1.

Source: Mathematics

DOI: 10.3390/math13223656

Authors: Seong Hyeon Noh, Jae Won Cho

Institution: Sejong University

Language: English

Topic: Multimodal Machine Learning Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $3,137

Funders: Ministry of Science and ICT, South Korea, Institute for Information and Communications Technology Promotion