Quantum Machine Learning

Quantum machine learning is the study of how quantum computers can be used to learn from data. It ranges from running classical models on quantum-inspired hardware to encoding data into quantum states and measuring observables as predictions.

The central promise is that quantum access to exponentially large Hilbert spaces might offer advantages for certain feature mappings, kernel methods, and generative modeling tasks — but only if we can overcome noise and limited qubit counts.

At Identity Lab we focus on variational and kernel-based approaches that are plausible on near-term hardware, and we benchmark them honestly against strong classical baselines ^[1]J. Preskill (2018). Quantum computing in the NISQ era and beyond. Quantum 2, 79. ^[2]J. Biamonte et al. (2017). Quantum machine learning. Nature 549, 195–202..

The field has grown rapidly. Today's QML toolbox includes variational circuits, quantum convolutional networks, generative adversarial models, and quantum kernels ^[3]V. Havlíček et al. (2019). Supervised learning with quantum-enhanced feature spaces. Nature 567, 209–212. ^[4]K. Mitarai et al. (2018). Quantum circuit learning. Phys. Rev. A 98, 032309..

Current Approaches

• •Variational Quantum Classifiers (VQC) — train parameterized circuits with classical optimizers ^[5]E. Farhi & H. Neven (2018). Classification with quantum neural networks on near term processors. arXiv:1802.06002..
• •Quantum Convolutional Neural Networks — hierarchical circuits inspired by tensor networks ^[6]I. Cong et al. (2019). Quantum convolutional neural networks. Nat. Phys. 15, 1273–1278..
• •Quantum GANs — quantum generator or discriminator circuits for generative modeling ^[7]S. Lloyd & C. Weedbrook (2018). Quantum generative adversarial learning. Phys. Rev. Lett. 121, 040502. ^[8]P.-L. Dallaire-Demers & N. Killoran (2018). Quantum generative adversarial networks. Phys. Rev. A 98, 012324..
• •Quantum Kernels — use quantum feature maps to compute similarity measures in high-dimensional Hilbert spaces ^[9]M. H. Amin et al. (2018). Quantum Boltzmann machine. Phys. Rev. X 8, 021050. ^[10]M. Schuld & N. Killoran (2019). Quantum machine learning in feature Hilbert spaces. Phys. Rev. Lett. 122, 040504..

Key Challenges

• •Barren plateaus — gradients can become exponentially small in deep circuits, making training hard ^[11]J. R. McClean et al. (2018). Barren plateaus in quantum neural network training landscapes. Nat. Commun. 9, 4812. ^[12]S. Wang et al. (2021). Noise-induced barren plateaus in variational quantum algorithms. Nat. Commun. 12, 1..
• •Noise — near-term devices are noisy, and error mitigation adds overhead ^[12]S. Wang et al. (2021). Noise-induced barren plateaus in variational quantum algorithms. Nat. Commun. 12, 1..
• •Data loading — encoding classical data into quantum states can dominate the computational cost ^[13]M. Schuld et al. (2021). Effect of data encoding on the expressive power of variational quantum-machine-learning models. Phys. Rev. A 103, 032430. ^[14]R. LaRose & B. Coyle (2020). Robust data encodings for quantum classifiers. Phys. Rev. A 102, 032420..