{"id":"arxiv/2607.00961","url":"https://arxiv.org/abs/2607.00961","tags":["quant-ph","cs.LG"],"score":0,"title":"Bridging Quantum Computing Paradigms toward Semiconductor Yield: A Controlled CV-versus-DV Comparison on Wafer-Map Defect Classification","domain":"quant-ph","source":"arxiv","authors":["Yeonhong Kim","Jonghyeok Im","Monu Nath Baitha","Kyoungsik Kim"],"reviews":0,"subject":"quant-ph","summary":"Realizing quantum neural networks (QNNs) in industry requires knowing which quantum computing paradigm suits which task. Motivated by AI accelerators and high-bandwidth memory, where die stacking makes wafer-level defect screening central to yield, we study WM-811K wafer-map defect classification (eight classes), comparing the dominant paradigms, continuous-variable (CV) and discrete-variable (DV), under controlled conditions. To isolate the quantum circuit as the sole variable, a shared convolutional backbone (~4.3M parameters) feeds interchangeable heads (classical dense, CV-QNN, or…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-07-01","reactions":{}}