将量子电路作为模块嵌入残差网络，提升神经解码精度与表征质量。
Quantum-Inspired Hybrid Neural Networks for Neural Decoding: A Controlled Ablation Study of Learnable Quantum Sidecar Integration
- 在ResNet-50中加入可学习的量子侧车模块，通过梯度优化投影层。
- 模型在31类手写识别任务中平均提升0.19%准确率，表征结构发生本质变化。
- 测量引导训练改善表征几何，适合研究量子-神经混合架构的开发者。
本文研究在ResNet-50主干网络中集成参数化量子电路（PQC）作为残差侧车模块，用于31类神经种群解码——即从多神经元放电序列中重构想象书写的分类任务。在严格控制条件下（固定数据划分、随机种子和优化器），对比四种模型变体：基线模型、输入投影冻结的量子侧车、投影由主干梯度训练的量子侧车，以及根据测量结果引导角度编码的变体。使用4量子比特无噪声态向量模拟，结果表明：投影梯度训练的变体在4个种子中3个提升准确率（均值+0.19%，95%置信区间[-1.10%, +1.48%]），且4个种子均显著降低线性CKA相似度（Δ=-0.025），表明表征结构发生真实重组。九种变体的消融实验表明，浅层简单架构最有效且可复现。测量引导训练始终改善表征几何，但不牺牲准确性。所有实验均基于当前近中期超导硬件约束的模拟，未宣称量子计算优势。
原文摘要 · Abstract (English)
We study parameterized quantum circuits (PQCs) integrated as residual sidecar modules within a ResNet-50 backbone for 31-class neural population decoding---imagined handwriting classification from multi-neuron spike rasters. Under strictly controlled conditions (fixed data splits, seeds, and optimizer), we compare four model variants: baseline, quantum sidecar with frozen input projection, quantum sidecar with backbone-gradient-trained projection, and a measurement-guided variant that aligns angle encodings with circuit measurement outcomes. The backbone-gradient variant improves accuracy in 3/4 seeds (+0.19% mean, 95% CI [-1.10%, +1.48%]) and consistently reduces Linear CKA similarity to baseline features ($Δ=-0.025$, 4/4 seeds), indicating genuine structural reorganization of representations. A nine-variant ablation identifies simple shallow architectures as the most effective and reproducible configuration. Measurement-guided training consistently improves representation geometry without reducing accuracy. All results use noiseless statevector simulation on 4 qubits, a regime chosen to reflect the practical constraints of current near-term superconducting hardware; no quantum computational advantage over classical methods is claimed.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。