Qiskit Machine Learning: an open-source library for quantum machine learning tasks at scale on quantum hardware and classical simulators

TL;DR

Qiskit ML presents an open-source, modular library that unifies quantum circuit design, simulation, and hardware execution for quantum machine learning tasks, enabling experiments on both classical simulators and quantum hardware. Built atop Qiskit primitives, it supports a broad algorithmic spectrum including fidelity-based quantum kernels, quantum neural networks, variational classifiers/regressors, and quantum SVM variants, with gradient estimation approaches like the Parameter-Shift rule $\partial f(\theta)/\partial \theta_i$ and SPSA for noisy devices. The framework emphasizes hardware-aware execution via IBM Runtime and Aer backends, integrates with Python data science stacks, and provides extensibility for researchers to add new algorithms and optimizers (e.g., L-BFGS-B, COBYLA, SLSQP, SPSA, ADAM, NFT). It further surveys literature leveraging Qiskit ML to advance QML on real hardware and simulators, highlighting progress in noise resilience, scalability, and practical applications, and invites community involvement under the Apache 2.0 license to sustain rapid development. Overall, the work demonstrates a practical pathway to scale quantum-enhanced machine learning from near-term devices toward fault-tolerant architectures, fostering reproducible workflows and broad adoption across disciplines.

Abstract

We present Qiskit Machine Learning (ML), a high-level Python library that combines elements of quantum computing with traditional machine learning. The API abstracts Qiskit's primitives to facilitate interactions with classical simulators and quantum hardware. Qiskit ML started as a proof-of-concept code in 2019 and has since been developed to be a modular, intuitive tool for non-specialist users while allowing extensibility and fine-tuning controls for quantum computational scientists and developers. The library is available as a public, open-source tool and is distributed under the Apache version 2.0 license.

Figures (1)

• Figure 1: UML diagram of the Qiskit ML library, showing the class hierarchy of the core components, algorithms and dependencies. The diagram categorises machine learning approaches into kernel-based, neural network-based, and Bayesian methods, linking them to their respective problem classes: classification and regression. It highlights key elements such as fidelity quantum kernels, trainable kernels, quantum neural networks, and quantum support vector machines, structured under core algorithms and supported by the Qiskit primitives ( Sampler and Estimator).