Beyond Barren Plateaus: A Scalable Quantum Convolutional Architecture for High-Fidelity Image Classification

TL;DR

This paper introduces a scalable quantum convolutional neural network architecture that effectively mitigates barren plateaus, enabling high-accuracy image classification on MNIST with fewer parameters than classical CNNs.

Contribution

The authors propose a novel QCNN design with localized cost functions and tensor-network initialization, significantly improving training efficiency and accuracy.

Findings

Achieves 98.7% accuracy on MNIST, surpassing baseline QCNNs.

Requires logarithmic fewer parameters than classical CNNs for similar performance.

Provides empirical evidence of mitigating barren plateaus in quantum neural networks.

Abstract

While Quantum Convolutional Neural Networks (QCNNs) offer a theoretical paradigm for quantum machine learning, their practical implementation is severely bottlenecked by barren plateaus -- the exponential vanishing of gradients -- and poor empirical accuracy compared to classical counterparts. In this work, we propose a novel QCNN architecture utilizing localized cost functions and a hardware-efficient tensor-network initialization strategy to provably mitigate barren plateaus. We evaluate our scalable QCNN on the MNIST dataset, demonstrating a significant performance leap. By resolving the gradient vanishing issue, our optimized QCNN achieves a classification accuracy of 98.7\%, a substantial improvement over the baseline QCNN accuracy of 52.32\% found in unmitigated models. Furthermore, we provide empirical evidence of a parameter-efficiency advantage, requiring $\mathcal{O}(\log N)$ fewer trainable parameters than equivalent classical CNNs to achieve $>95\%$ convergence. This work bridges the gap between theoretical quantum utility and practical application, providing a scalable framework for quantum computer vision tasks without succumbing to loss landscape concentration.