Quantum Support Vector Machine
Also known as: QSVM, quantum kernel
Quantum Support Vector Machine (QSVM) is a quantum machine learning algorithm combining quantum feature spaces with classical SVM training. Proposed by Rebentrost et al. in 2014, QSVM leverages quantum processors to compute kernel functions, potentially offering speedup for classification problems while remaining practical on near-term quantum devices.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
QSVM is used for binary classification when quantum feature spaces can separate data more efficiently. It requires quantum access to data and quantum computers with sufficient coherence.
Strengths & limitations
- Quantum feature space can be exponentially large, enabling separation of classes classically inseparable.
- Combines power of quantum circuits with robustness of SVM.
- Implementable on near-term quantum devices (hybrid quantum-classical).
- Demonstrated on real quantum hardware (IBM, Rigetti).
- Kernel-based approach separates feature mapping from optimization.
- Quantum speedup not proven for most realistic datasets.
- Noise limits quantum kernel evaluation accuracy.
- Encoding and readout may require exponential resources in worst case.
- Classical SVM can often match or exceed quantum performance via kernel trick.
- Scalability to large datasets limited by quantum coherence.
Frequently asked
What is a quantum kernel and how is it computed?
A quantum kernel is the inner product ⟨ψ(x_i)|ψ(x_j)⟩ between two quantum-encoded states. It is computed by applying the inverse of the feature encoding circuit to create a SWAP test, then measuring.
Can QSVM provide exponential speedup?
Theoretical speedups have been claimed for certain problem structures (e.g., strongly correlated data). However, proving speedup for realistic datasets remains open. Most empirical results show quantum and classical SVMs perform similarly.
How is the feature encoding chosen?
Feature encoding is often problem-specific. Common choices include amplitude encoding (mapping data to quantum amplitudes) or angle encoding (mapping to rotation angles). Ansatz design significantly affects performance.
What is the SWAP test and why is it used?
The SWAP test measures the overlap of two quantum states. A circuit applies controlled-SWAP between states, then measures an ancilla qubit. Overlap probability is extracted from measurement statistics.
How does QSVM compare to classical neural networks?
QSVM leverages quantum feature spaces; neural networks learn features via training. Neural networks are more flexible and scalable; QSVM offers potential quantum advantage if quantum features are exponentially separable, but this is hard to guarantee.
Sources
- Rebentrost, P., Mohseni, M., Lloyd, S. (2014). Quantum support vector machine for big data classification. Physical Review Letters, 113, 130503. DOI: 10.1103/PhysRevLett.113.130503 ↗
- Havlíček, V., Córcoles, A. D., Temme, K., et al. (2019). Supervised learning with quantum-enhanced feature spaces. Nature, 567, 209–212. DOI: 10.1038/s41586-019-0980-2 ↗
- Liu, Y., Arunachalam, S., Temme, K. (2021). A rigorous and robust quantum speed-up in supervised machine learning. arXiv preprint arXiv:2010.07471. link ↗
How to cite this page
ScholarGate. (2026, June 3). Quantum Support Vector Machine. ScholarGate. https://scholargate.app/en/quantum-computing/quantum-svm
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Quantum Approximate Optimization AlgorithmQuantum Computing↔ compare
- Variational Quantum EigensolverQuantum Computing↔ compare