{"id":"arxiv/2602.16097","url":"https://arxiv.org/abs/2602.16097","tags":["quant-ph"],"score":0,"title":"Local and Multi-Scale Strategies to Mitigate Exponential Concentration in Quantum Kernels","domain":"quant-ph","source":"arxiv","authors":["Claudia Zendejas-Morales","Debashis Saikia","Utkarsh Singh"],"reviews":0,"subject":"quant-ph","summary":"Fidelity-based quantum kernels provide a direct interface between quantum feature maps and classical kernel methods, but they can exhibit exponential concentration: with increasing system size or circuit expressivity, the Gram matrix approaches the identity and suppresses informative similarity structure. We present an empirical study of two mitigation strategies implemented in Qiskit: (i) local (patch-wise) kernels that aggregate subsystem similarities, and (ii) multi-scale kernels that mix local and global similarity across patch granularities. We benchmark baseline, local, and multi-scale…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-02-18","reactions":{}}