A Pauli-transfer-matrix analysis of QELMs reveals the full set of nonlinear Pauli features generated by encoding and transformed by quantum channels, producing an interpretable classical nonlinear vector autoregression model that approximates flow maps in dynamical systems.
hub
Kornjača, H.-Y
15 Pith papers cite this work, alongside 9 external citations. Polarity classification is still indexing.
15
Pith papers citing it
9
external citations · Pith
abstract
Quantum machine learning has gained considerable attention as quantum technology advances, presenting a promising approach for efficiently learning complex data patterns. Despite this promise, most contemporary quantum methods require significant resources for variational parameter optimization and face issues with vanishing gradients, leading to experiments that are either limited in scale or lack potential for quantum advantage. To address this, we develop a general-purpose, gradient-free, and scalable quantum reservoir learning algorithm that harnesses the quantum dynamics of neutral-atom analog quantum computers to process data. We experimentally implement the algorithm, achieving competitive performance across various categories of machine learning tasks, including binary and multi-class classification, as well as timeseries prediction. Effective and improving learning is observed with increasing system sizes of up to 108 qubits, demonstrating the largest quantum machine learning experiment to date. We further observe comparative quantum kernel advantage in learning tasks by constructing synthetic datasets based on the geometric differences between generated quantum and classical data kernels. Our findings demonstrate the potential of utilizing classically intractable quantum correlations for effective machine learning. We expect these results to stimulate further extensions to different quantum hardware and machine learning paradigms, including early fault-tolerant hardware and generative machine learning tasks.
hub tools
citation-role summary
background 3
citation-polarity summary
fields
quant-ph 15roles
background 3polarities
background 3representative citing papers
A quantum reservoir network using GHZ-state preparation achieves an order-of-magnitude RMSE improvement over prior QRN designs on latent-space prediction of the Kuramoto-Sivashinsky equation.
Hybrid quantum-classical reservoir computing enables nonlinear temporal processing of quantum states and outperforms pure quantum or classical reservoirs in both full-tomography and single-axis measurement regimes.
Proposes a hybrid quantum-classical framework for running event-based graph neural networks on neutral-atom processors by mapping events to atoms and programming the Rydberg Hamiltonian to realize message passing.
Develops an invariant-based framework connecting Pauli Lie algebras to transvection-generated Clifford subgroups for quantum reachability and dynamics analysis.
Introduces tunable partial-SWAP for controllable memory capacity in quantum reservoir networks, modeled as controlled amplitude-damping and validated via STMC and NARMA-5 benchmarks on simulators and IBM QPUs.
Generalized Krylov complexity predicts the minimum time to realize target operations in analog quantum simulators such as Rydberg atom arrays.
A fixed time-encoding, not barren-plateau gradient vanishing, sets a 33-dimensional capacity ceiling (11 accessible Fourier modes per observable) that prevents a four-qubit variational QPINN from fitting Lorenz-63.
Recurrent quantum feature maps achieve lower mean squared error than echo state networks and multilayer perceptrons on Mackey-Glass prediction using compact quantum circuits.
Indirect measurements in quantum reservoir computing improve execution time scaling, overall performance, and memory capacity over projective measurements and classical feedback methods.
Jaynes-Cummings qubit-boson systems show superior nonlinear memory capacity and comparable Mackey-Glass forecasting performance when used as quantum reservoirs.
A quantum echo-state network is implemented on NISQ superconducting qubits and shown to predict long chaotic trajectories from the Lorenz system with memory persisting over 100 times the median T1/T2 time.
Quantum feature surrogates let a quantum processor act as a teacher on a small data subsample while a classical surrogate applies the learned representations to the entire industrial dataset.
A parametrized four-qubit digital QRC model with ridge-regression readout matches the classical Prophet benchmark on dynamic time warping for ATM cash-demand forecasting but underperforms on MAE and NMSE across noiseless simulation, noise emulation, and real IQM hardware.
TDVP-MPS simulations of Rydberg atom chains mitigate exponential concentration in QELM, yielding competitive MNIST accuracy via controlled entanglement and disorder without requiring exact quantum dynamics.
citing papers explorer
-
Theory and interpretability of Quantum Extreme Learning Machines: a Pauli-transfer matrix approach
A Pauli-transfer-matrix analysis of QELMs reveals the full set of nonlinear Pauli features generated by encoding and transformed by quantum channels, producing an interpretable classical nonlinear vector autoregression model that approximates flow maps in dynamical systems.
-
Leveraging Metrologically Useful States in Quantum Reservoir Networks
A quantum reservoir network using GHZ-state preparation achieves an order-of-magnitude RMSE improvement over prior QRN designs on latent-space prediction of the Kuramoto-Sivashinsky equation.
-
Temporal processing of quantum states with hybrid quantum-classical reservoirs
Hybrid quantum-classical reservoir computing enables nonlinear temporal processing of quantum states and outperforms pure quantum or classical reservoirs in both full-tomography and single-axis measurement regimes.
-
Analog Quantum Asynchronous Event-Based Graph Neural Network
Proposes a hybrid quantum-classical framework for running event-based graph neural networks on neutral-atom processors by mapping events to atoms and programming the Rydberg Hamiltonian to realize message passing.
-
From Pauli Strings to Quantum Dynamics: A Unified Characterization
Develops an invariant-based framework connecting Pauli Lie algebras to transvection-generated Clifford subgroups for quantum reachability and dynamics analysis.
-
Controllable Quantum Memory Capacity in Quantum Reservoir Networks with Tunable partial-SWAPs
Introduces tunable partial-SWAP for controllable memory capacity in quantum reservoir networks, modeled as controlled amplitude-damping and validated via STMC and NARMA-5 benchmarks on simulators and IBM QPUs.
-
Bridging Krylov Complexity and Universal Analog Quantum Simulator
Generalized Krylov complexity predicts the minimum time to realize target operations in analog quantum simulators such as Rydberg atom arrays.
-
An architectural capacity ceiling, not a barren plateau: why a fixed-encoding variational quantum circuit cannot fit the Lorenz-63 attractor
A fixed time-encoding, not barren-plateau gradient vanishing, sets a 33-dimensional capacity ceiling (11 accessible Fourier modes per observable) that prevents a four-qubit variational QPINN from fitting Lorenz-63.
-
Recurrent Quantum Feature Maps for Reservoir Computing
Recurrent quantum feature maps achieve lower mean squared error than echo state networks and multilayer perceptrons on Mackey-Glass prediction using compact quantum circuits.
-
Harnessing quantum back-action for time-series processing
Indirect measurements in quantum reservoir computing improve execution time scaling, overall performance, and memory capacity over projective measurements and classical feedback methods.
-
Quantum reservoir computing in Jaynes-Cummings models: Nonlinear memory and time-series prediction
Jaynes-Cummings qubit-boson systems show superior nonlinear memory capacity and comparable Mackey-Glass forecasting performance when used as quantum reservoirs.
-
Quantum Observers: A NISQ Hardware Demonstration of Chaotic State Prediction Using Quantum Echo-state Networks
A quantum echo-state network is implemented on NISQ superconducting qubits and shown to predict long chaotic trajectories from the Lorenz system with memory persisting over 100 times the median T1/T2 time.
-
Off-line quantum-advantage feature extraction for industrial production
Quantum feature surrogates let a quantum processor act as a teacher on a small data subsample while a classical surrogate applies the learned representations to the entire industrial dataset.
-
Digital Quantum Reservoir Computing for ATM Time Series Prediction
A parametrized four-qubit digital QRC model with ridge-regression readout matches the classical Prophet benchmark on dynamic time warping for ATM cash-demand forecasting but underperforms on MAE and NMSE across noiseless simulation, noise emulation, and real IQM hardware.
-
Harnessing Quantum Dynamics for Robust and Scalable Quantum Extreme Learning Machines
TDVP-MPS simulations of Rydberg atom chains mitigate exponential concentration in QELM, yielding competitive MNIST accuracy via controlled entanglement and disorder without requiring exact quantum dynamics.