![[2602.16233] DistributedEstimator: Distributed Training of Quantum Neural Networks via Circuit Cutting](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.16233] DistributedEstimator: Distributed Training of Quantum Neural Networks via Circuit Cutting
Summary
The paper presents a novel approach to distributed training of quantum neural networks using circuit cutting, addressing overheads and performance implications in iterative training pipelines.
Why It Matters
As quantum computing advances, optimizing the training of quantum neural networks becomes crucial. This research provides insights into the efficiency and scalability of distributed training methods, which can significantly impact the development of practical quantum machine learning applications.
Key Takeaways
- Circuit cutting can introduce significant overheads in distributed quantum training.
- The study quantifies the impact of cutting on accuracy and robustness in learning workloads.
- Effective scheduling and reconstruction strategies are essential for optimizing performance.
Computer Science > Distributed, Parallel, and Cluster Computing arXiv:2602.16233 (cs) [Submitted on 18 Feb 2026] Title:DistributedEstimator: Distributed Training of Quantum Neural Networks via Circuit Cutting Authors:Prabhjot Singh, Adel N. Toosi, Rajkumar Buyya View a PDF of the paper titled DistributedEstimator: Distributed Training of Quantum Neural Networks via Circuit Cutting, by Prabhjot Singh and 1 other authors View PDF HTML (experimental) Abstract:Circuit cutting decomposes a large quantum circuit into a collection of smaller subcircuits. The outputs of these subcircuits are then classically reconstructed to recover the original expectation values. While prior work characterises cutting overhead largely in terms of subcircuit counts and sampling complexity, its end-to-end impact on iterative, estimator-driven training pipelines remains insufficiently measured from a systems perspective. In this paper, we propose a cut-aware estimator execution pipeline that treats circuit cutting as a staged distributed workload and instruments each estimator query into partitioning, subexperiment generation, parallel execution, and classical reconstruction phases. Using logged runtime traces and learning outcomes on two binary classification workloads (Iris and MNIST), we quantify cutting overheads, scaling limits, and sensitivity to injected stragglers, and we evaluate whether accuracy and robustness are preserved under matched training budgets. Our measurements show that cuttin...