PQFA: Parallel Quantum Feature Augmentation of Fused Representations for Multimodal Classification

Parallel Quantum Feature Augmentation (PQFA) is a hybrid quantum-classical framework designed for multimodal classification. It focuses on post-fusion enhancement by applying multiple shallow variational quantum circuits to fused features. The system utilizes frozen RoBERTa and ViT encoders to extract text and image representations, which are processed via bidirectional cross-attention, attentive pooling, and adaptive gated fusion. These fused features are amplitude-encoded into parallel quantum circuits, and the resulting measurement readouts are concatenated with classical representations for prediction.