Tackling Sampling Noise in Physical Systems for Machine Learning Applications: Fundamental Limits and Eigentasks

Abstract

The expressive capacity of physical systems employed for learning is limited by the unavoidable presence of noise in their extracted outputs. Though present in physical systems across both the classical and quantum regimes, the precise impact of noise on learning remains poorly understood. Focusing on supervised learning, we present a mathematical framework for evaluating the resolvable expressive capacity (REC) of general physical systems under finite sampling noise, and provide a methodology for extracting its extrema, the eigentasks. Eigentasks are a native set of functions that a given physical system can approximate with minimal error. We show that the REC of a quantum system is limited by the fundamental theory of quantum measurement, and obtain a tight upper bound for the REC of any finitely-sampled physical system. We then provide empirical evidence that extracting low-noise eigentasks can lead to improved performance for machine learning tasks such as classification, displaying robustness to overfitting. We present analyses suggesting that correlations in the measured quantum system enhance learning capacity by reducing noise in eigentasks. The applicability of these results in practice is demonstrated with experiments on superconducting quantum processors. Our findings have broad implications for quantum machine learning and sensing applications.

arXiv comment37 pages, 10 figures, 1 table. Significantly expanded version of arXiv:2301.00042 [quant-ph], covering physical systems operating in the classical and quantum regimes

References in corpus (31)

• Variational Quantum Algorithms
• Deep Learning with Coherent Nanophotonic Circuits
• All-Optical Machine Learning Using Diffractive Deep Neural Networks
• Supervised learning with quantum enhanced feature spaces
• Quantum metrology
• Recent Advances in Physical Reservoir Computing: A Review
• The Role of Relative Entropy in Quantum Information Theory
• Expressibility and entangling capability of parameterized quantum circuits for hybrid quantum-classical algorithms
• Unified View of Quantum and Classical Correlations
• The effect of data encoding on the expressive power of variational quantum machine learning models
• Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems
• Connecting ansatz expressibility to gradient magnitudes and barren plateaus
• Procedure for systematically tuning up crosstalk in the cross resonance gate
• Experimentally realized in situ backpropagation for deep learning in nanophotonic neural networks
• Random Quantum Circuits are Approximate 2-designs
• Harnessing disordered quantum dynamics for machine learning
• Opportunities in Quantum Reservoir Computing and Extreme Learning Machines
• Optical Reservoir Computing using multiple light scattering for chaotic systems prediction
• Fisher Information in Noisy Intermediate-Scale Quantum Applications
• Dynamical phase transitions in quantum reservoir computing
• Temporal Information Processing on Noisy Quantum Computers
• Spectral Bias and Task-Model Alignment Explain Generalization in Kernel Regression and Infinitely Wide Neural Networks
• Photonic extreme learning machine by free-space optical propagation
• Rapid Time Series Prediction with a Hardware-Based Reservoir Computer
• Symbolic integration with respect to the Haar measure on the unitary group
• Information Processing Capacity of Spin-Based Quantum Reservoir Computing Systems
• Potential and limitations of quantum extreme learning machines
• Scalable photonic platform for real-time quantum reservoir computing
• Expressivity of Quantum Neural Networks
• Self-learning Machines based on Hamiltonian Echo Backpropagation
• The Computational Capacity of LRC, Memristive and Hybrid Reservoirs

Cited by in corpus (14)

• Dissipation as a resource for Quantum Reservoir Computing
• Microwave signal processing using an analog quantum reservoir computer
• Overcoming the Coherence Time Barrier in Quantum Machine Learning on Temporal Data
• Role of coherence in many-body Quantum Reservoir Computing
• Retrieving past quantum features with deep hybrid classical-quantum reservoir computing
• Practical Few-Atom Quantum Reservoir Computing
• Stochastic Reservoir Computers
• Robust Quantum Reservoir Computing for Molecular Property Prediction
• Practical Trainable Temporal Postprocessor for Multistate Quantum Measurement
• Pitfalls when tackling the exponential concentration of parameterized quantum models
• Restrictions on Physical Stochastic Reservoir Computers
• Edge of Many-Body Quantum Chaos in Quantum Reservoir Computing
• Practical Quantum Reservoir Computing in Rydberg Atom Arrays
• Connection between memory performance and optical absorption in quantum reservoir computing