How Much Reconstruction Does Quantum Machine Learning Need? Late Fusion of Independently Trained Quantum Subcircuits
Prabhjot Singh, Adel N. Toosi, Rajkumar Buyya
quant-ph, cs.DC, cs.LG
Submitted: 2026-08-06
Comments: 11 pages, 6 figures. Includes technical appendix
License: http://creativecommons.org/licenses/by/4.0/
The gist: Circuit cutting lets a large quantum neural network (QNN) run as independent subcircuits on small devices, but rebuilding its outputs by reconstruction carries a classical sampling overhead
Terminology
Abstract
Circuit cutting lets a large quantum neural network (QNN) run as independent subcircuits on small devices, but rebuilding its outputs by reconstruction carries a classical sampling overhead exponential in the number of cuts - the dominant runtime cost in prior work. We ask whether, for machine-learning tasks, this step is necessary, and replace it with late fusion: each subcircuit is trained and measured independently, and a small classical head combines their outputs - a linear-cost, decision-level combination borrowed from multimodal learning. To characterize the trade-off we introduce a quantumness dial Q, a tunable reconstruction budget interpolating from pure fusion to full reconstruction, and a cut-entanglement diagnostic that indicates how much reconstruction a task needs (Spearman rho=0.59 over 104 runs). Across synthetic and standard datasets, independently trained late fusion matches full reconstruction accuracy within 0.04 at every point of the controlled sweep and on every classical benchmark, at exponentially lower cost; it is also markedly more robust to shot and device noise. Controlled entangled-data experiments locate the boundary where fusion must fail. We do not claim advantage over classical machine learning - consistent with recent benchmarking, quantum offers no accuracy edge on these datasets. Late fusion is thus an efficient, noise-robust, self-characterizing alternative to reconstruction for circuit-cutting QML.
Sources
- On quantum ensembles of quantum classifiers
- Multimodal Machine Learning: A Survey and Taxonomy
- Better than classical? The subtle art of benchmarking quantum machine learning models
- Optimal wire cutting with classical communication
- Circuit Knitting Faces Exponential Sampling Overhead Scaling Bounded by Entanglement Cost
- Distributed Quantum Neural Networks via Partitioned Features Encoding
- Ensemble-learning error mitigation for variational quantum shallow-circuit classifiers
- Fast quantum circuit cutting with randomized measurements
- Cutting is All You Need: Execution of Large-Scale Quantum Neural Networks on Limited-Qubit Devices
- High Dimensional Quantum Machine Learning With Small Quantum Computers
- Classical Ensembles of Single-Qubit Quantum Variational Circuits for Classification
- Constructing a virtual two-qubit gate by sampling single-qubit operations
- Expressive and Scalable Quantum Fusion for Multimodal Learning
- Simulating Large Quantum Circuits on a Small Quantum Computer
- Circuit knitting with classical communication
- Quantum ensembles of quantum classifiers
- DistributedEstimator: Distributed Training of Quantum Neural Networks via Circuit Cutting
- CutQC: Using Small Quantum Computers for Large Quantum Circuit Evaluations
- Optimal joint cutting of two-qubit rotation gates
- Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
Related papers
- Reconquering Bell sampling on qudits: stabilizer learning and testing, quantum pseudorandomness bounds, and more
- Encrypted clones can leak: Classification of informative subsets in Quantum Encrypted Cloning
- Pure states are optimal for total energy-constrained quantum channel discrimination
- Polynomial-time classical and quantum simulation of quantum impurity models
- Incoherent Imaging with Spatially Structured Quantum Probes
- A quasiparticle-protected superconducting qubit