Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2605.30429

Attention-based optimizer for symmetry finding

Shreya Banerjee, Vinodh Raj Rajagopal Muthu, Charlie Nation, Rick P. A. Simon, Francesco Martini, Alessandro Ricottone, Federico Cerisola, Luca Dellantonio

quant-phcs.LGSubmitted May 28, 2026• Announced Jun 1, 2026• 0 citations
arXivHTMLPDF

Abstract

Finding symmetries is crucial for understanding physical models. In this work, we present an optimization framework that searches Pauli symmetries of Hamiltonians, merging the fields of machine learning with automated symmetry finding. Built on a Set-Transformer architecture, our framework uses self-attention to encode the pairwise and higher-order correlations among the Pauli-Strings. The relations are then decoded as a candidate, which is further optimized with a custom commutation-based objective, and mapped to a symmetry of the input Hamiltonian. We apply our method to random Pauli Hamiltonians, periodic one and two dimensional transverse-field Ising model and the Toric code. We show that for physical Hamiltonians (Ising and Toric), our framework succeeds with near-deterministic probability while providing substantial advantage compared to state-of-the-art strategies. For random Pauli Hamiltonians, we estimate the required computational resources, specifically the number of parallel starts and the number of GPUs, to find a symmetry with high success probability under fixed design specifications.

Comments
9+4 pages, 2 Figures, Comments welcome

Related papers

Symmetry Constraints Regularize Neural Quantum State Learning

Turbasu Chatterjee, Manas Sajjan, Songbo Xie +4 more

Neural quantum states (NQS) offer highly expressive variational wavefunctions, but their optimization is frequently bottlenecked by redundant parameters and poorly conditioned landscapes. We demonstrate that embedding Hamiltonian symmetries directly into the variational parameterization geometrically regularizes this learning problem. For Boltzmann-family NQS, we enforce symmetries by tying local Pauli-$Z$ generators along physical geometric orbits, analytically collapsing the trainable coefficient space prior to optimization. To quantify the resulting optimization geometry, we introduce a geometric metric built on the Jacobian and Hessian of the optimization landscape. This framework evaluates the fraction of the physically accessible state space that corresponds to high-quality, low-energy solutions. Evaluating our approach on transverse-field Ising (TFIM) and XXZ spin chains shows that symmetry compilation excises the vast majority of parameters while maintaining ground-state accuracy within the resolution of the reported benchmarks. In large TFIM systems, strong spatial constraints compress thousands of parameters down to tens, delivering substantial runtime accelerations. Our geometric diagnostics indicate that symmetry produces a more favorable target-aware geometry by concentrating the reachable state space around low-energy solutions while retaining broad target basins. Together, our results indicate that symmetry compilation concentrates the expressive power of NQS on states relevant to the target problem, thereby reducing model size and training cost without sacrificing accuracy.

quant-pharXiv:2608.08798•Aug 11, 2026•0 citations

Enhancing Pauli Correlation Encoding for quantum optimization via systematic expressivity analysis

Riku Usuki, Don Arai, Ken N. Okada +1 more

Quantum approaches for combinatorial optimization problems have attracted considerable attention in recent years. Among these approaches, Pauli Correlation Encoding (PCE) has emerged as a promising framework for quantum devices with limited qubit resources because it embeds optimization variables in expectation values of Pauli strings. However, the mechanisms underlying its performance and the reasons for its saturation remain unclear. In this work, we investigate these questions through a systematic analysis of expressivity and trainability. First, we compare PCE with classical surrogate models based on tensor networks whose structures progressively approach the topology of the PCE circuit. The results show that PCE attains comparable solution quality with substantially fewer trainable parameters, indicating strong parameter efficiency. Second, to determine whether the performance saturation of conventional PCE is caused by insufficient expressivity or by optimization difficulty, we perform a diagnostic expressivity test in which the circuit is trained toward reference configurations for Max-Cut. The results show that even shallow PCE circuits can represent strong solutions, indicating that the main bottleneck is not the representational power of the ansatz, but the trainability under the relaxed objective function. Motivated by this finding, we propose a multistage continuation framework that gradually transforms a smooth relaxed objective into a sharper objective that more closely approximates the target discrete problem. Numerical experiments on G-set instances with 800 vertices show that the proposed method consistently outperforms conventional PCE and is competitive with representative graph neural network (GNN) methods. These results clarify the main factors behind PCE performance and provide a practical strategy for improving PCE on quantum devices with limited qubit resources.

quant-pharXiv:2609.09718•Sep 10, 2026•0 citations

Adding machine learning within Hamiltonians: Renormalization group transformations, symmetry breaking and restoration

Dimitrios Bachtis, Gert Aarts A, Biagio Lucini A

We present a physical interpretation of machine learning functions, opening up the possibility to control properties of statistical systems via the inclusion of these functions in Hamiltonians.In particular, we include the predictive function of a neural network, designed for phase classification, as a conjugate variable coupled to an external field within the Hamiltonian of a system.Results in the two-dimensional Ising model evidence that the field can induce an order-disorder phase transition by breaking or restoring the symmetry, in contrast with the field of the conventional order parameter which causes explicit symmetry breaking.The critical behavior is then studied by proposing a Hamiltonian-agnostic reweighting approach and forming a renormalization group mapping on quantities derived from the neural network.Accurate estimates of the critical point and of the critical exponents related to the operators that govern the divergence of the correlation length are provided.We conclude by discussing how the method provides an essential step toward bridging machine learning and physics.

arXiv:2010.00054•Physical Review Research•Feb 10, 2021•13 citations

TARE: Block Encoding Linear Combinations of Pauli Strings Without Ancilla State Preparation

Niclas Schillo, Andreas Sturm, Rüdiger Quay

Quantum algorithms based on Quantum Signal Processing (QSP) offer the potential for speedups across a broad range of applications, with block encodings serving as the central input model. In this framework, non-unitary matrices are embedded into larger unitary operators, and the cost of constructing these encodings often dominates the overall gate complexity. In this work, we introduce Tag-and-Restore Encoding (TARE), a block-encoding method for linear combinations of Pauli strings. In this method coefficient magnitudes are absorbed into a unitary built from a set of mutually anti-commuting Pauli strings acting on the system register. These Pauli strings are then mapped to the target Pauli strings through appropriate transformations, yielding a block encoding of the target operator. The ancilla register size scales logarithmically with the number of Pauli strings and can be extended to larger registers providing a width/depth tradeoff. We evaluate TARE through numerical simulations of the transverse-field Ising model, the Jordan-Wigner image of a fermionic star Hamiltonian, and random Pauli-string operators. Compared with standard Linear Combination of Unitaries (LCU), TARE substantially reduces the T-gate count while improving circuit depth in several cases. These results suggest that TARE can provide resource-efficient block encodings for a wide range of relevant systems.

quant-pharXiv:2601.05740•Jan 12, 2026•0 citations

Symmetries for a high-level neural decoder on the toric code

Thomas Wagner, Hermann Kampermann, Dagmar Bruß

Surface codes are a promising method of quantum error correction and the basis of many proposed quantum computation implementations. However, their efficient decoding is still not fully explored. Recently, approaches based on machine learning techniques have been proposed by Torlai and Melko [Phys. Rev. Lett. 119, 030501 (2017)] as well as Varsamopoulos et al. [Quantum Sci. Technol. 3, 015004 (2017)]. In these approaches, a so-called high-level decoder is used to post-correct an underlying decoder by correcting logical errors. A significant problem is that these methods require large amounts of training data even for relatively small code distances. The above-mentioned methods were tested on the rotated surface code which encodes one logical qubit. Here, we show that they are viable even for the toric surface code which encodes two logical qubits. Furthermore, we explain how symmetries of the toric code can be exploited to reduce the amount of training data that is required to obtain good decoding results. Finally, we compare different underlying decoders and show that the accuracy of high-level decoding noticeably depends on the quality of the underlying decoder in the realistic case of imperfect training.

arXiv:1910.01662•Physical Review A•Oct 26, 2020•18 citations

Clifford symmetries in quantum many-body systems

Charlie Nation, Rick P. A. Simon, Shreya Banerjee +4 more

Obtaining the symmetries of a model is a critical step towards developing an understanding and ultimately analytically or numerically solving the model. However, finding symmetries is generally extremely complicated, often being the result of insightful thinking. In this work, we complement human ingenuity with an algorithm. We leverage the classically efficient Clifford group to find symmetries for arbitrary many-body Hamiltonians via a graph representation. We demonstrate our method on random and physical Hamiltonians, with instances of up to one thousand qubits and demonstrate how our approach can provide deeper understanding of the model.

quant-pharXiv:2605.18966•May 20, 2026•0 citations
Finding symmetries is crucial for understanding physical models. In this work, we present an optimization framework that searches Pauli symmetries of Hamiltonians, merging the fields of machine learning with automated symmetry finding. Built on a Set-Transformer architecture, our framework uses self-attention to encode the pairwise and higher-order correlations among the Pauli-Strings. The relations are then decoded as a candidate, which is further optimized with a custom commutation-based objective, and mapped to a symmetry of the input Hamiltonian. We apply our method to random Pauli Hamiltonians, periodic one and two dimensional transverse-field Ising model and the Toric code. We show that for physical Hamiltonians (Ising and Toric), our framework succeeds with near-deterministic probability while providing substantial advantage compared to state-of-the-art strategies. For random Pauli Hamiltonians, we estimate the required computational resources, specifically the number of parallel starts and the number of GPUs, to find a symmetry with high success probability under fixed design specifications.
Neural quantum states (NQS) offer highly expressive variational wavefunctions, but their optimization is frequently bottlenecked by redundant parameters and poorly conditioned landscapes. We demonstrate that embedding Hamiltonian symmetries directly into the variational parameterization geometrically regularizes this learning problem. For Boltzmann-family NQS, we enforce symmetries by tying local Pauli-ZZZ generators along physical geometric orbits, analytically collapsing the trainable coefficient space prior to optimization. To quantify the resulting optimization geometry, we introduce a geometric metric built on the Jacobian and Hessian of the optimization landscape. This framework evaluates the fraction of the physically accessible state space that corresponds to high-quality, low-energy solutions. Evaluating our approach on transverse-field Ising (TFIM) and XXZ spin chains shows that symmetry compilation excises the vast majority of parameters while maintaining ground-state accuracy within the resolution of the reported benchmarks. In large TFIM systems, strong spatial constraints compress thousands of parameters down to tens, delivering substantial runtime accelerations. Our geometric diagnostics indicate that symmetry produces a more favorable target-aware geometry by concentrating the reachable state space around low-energy solutions while retaining broad target basins. Together, our results indicate that symmetry compilation concentrates the expressive power of NQS on states relevant to the target problem, thereby reducing model size and training cost without sacrificing accuracy.
Quantum approaches for combinatorial optimization problems have attracted considerable attention in recent years. Among these approaches, Pauli Correlation Encoding (PCE) has emerged as a promising framework for quantum devices with limited qubit resources because it embeds optimization variables in expectation values of Pauli strings. However, the mechanisms underlying its performance and the reasons for its saturation remain unclear. In this work, we investigate these questions through a systematic analysis of expressivity and trainability. First, we compare PCE with classical surrogate models based on tensor networks whose structures progressively approach the topology of the PCE circuit. The results show that PCE attains comparable solution quality with substantially fewer trainable parameters, indicating strong parameter efficiency. Second, to determine whether the performance saturation of conventional PCE is caused by insufficient expressivity or by optimization difficulty, we perform a diagnostic expressivity test in which the circuit is trained toward reference configurations for Max-Cut. The results show that even shallow PCE circuits can represent strong solutions, indicating that the main bottleneck is not the representational power of the ansatz, but the trainability under the relaxed objective function. Motivated by this finding, we propose a multistage continuation framework that gradually transforms a smooth relaxed objective into a sharper objective that more closely approximates the target discrete problem. Numerical experiments on G-set instances with 800 vertices show that the proposed method consistently outperforms conventional PCE and is competitive with representative graph neural network (GNN) methods. These results clarify the main factors behind PCE performance and provide a practical strategy for improving PCE on quantum devices with limited qubit resources.
We present a physical interpretation of machine learning functions, opening up the possibility to control properties of statistical systems via the inclusion of these functions in Hamiltonians.In particular, we include the predictive function of a neural network, designed for phase classification, as a conjugate variable coupled to an external field within the Hamiltonian of a system.Results in the two-dimensional Ising model evidence that the field can induce an order-disorder phase transition by breaking or restoring the symmetry, in contrast with the field of the conventional order parameter which causes explicit symmetry breaking.The critical behavior is then studied by proposing a Hamiltonian-agnostic reweighting approach and forming a renormalization group mapping on quantities derived from the neural network.Accurate estimates of the critical point and of the critical exponents related to the operators that govern the divergence of the correlation length are provided.We conclude by discussing how the method provides an essential step toward bridging machine learning and physics.
Quantum algorithms based on Quantum Signal Processing (QSP) offer the potential for speedups across a broad range of applications, with block encodings serving as the central input model. In this framework, non-unitary matrices are embedded into larger unitary operators, and the cost of constructing these encodings often dominates the overall gate complexity. In this work, we introduce Tag-and-Restore Encoding (TARE), a block-encoding method for linear combinations of Pauli strings. In this method coefficient magnitudes are absorbed into a unitary built from a set of mutually anti-commuting Pauli strings acting on the system register. These Pauli strings are then mapped to the target Pauli strings through appropriate transformations, yielding a block encoding of the target operator. The ancilla register size scales logarithmically with the number of Pauli strings and can be extended to larger registers providing a width/depth tradeoff. We evaluate TARE through numerical simulations of the transverse-field Ising model, the Jordan-Wigner image of a fermionic star Hamiltonian, and random Pauli-string operators. Compared with standard Linear Combination of Unitaries (LCU), TARE substantially reduces the T-gate count while improving circuit depth in several cases. These results suggest that TARE can provide resource-efficient block encodings for a wide range of relevant systems.
Surface codes are a promising method of quantum error correction and the basis of many proposed quantum computation implementations. However, their efficient decoding is still not fully explored. Recently, approaches based on machine learning techniques have been proposed by Torlai and Melko [Phys. Rev. Lett. 119, 030501 (2017)] as well as Varsamopoulos et al. [Quantum Sci. Technol. 3, 015004 (2017)]. In these approaches, a so-called high-level decoder is used to post-correct an underlying decoder by correcting logical errors. A significant problem is that these methods require large amounts of training data even for relatively small code distances. The above-mentioned methods were tested on the rotated surface code which encodes one logical qubit. Here, we show that they are viable even for the toric surface code which encodes two logical qubits. Furthermore, we explain how symmetries of the toric code can be exploited to reduce the amount of training data that is required to obtain good decoding results. Finally, we compare different underlying decoders and show that the accuracy of high-level decoding noticeably depends on the quality of the underlying decoder in the realistic case of imperfect training.
Obtaining the symmetries of a model is a critical step towards developing an understanding and ultimately analytically or numerically solving the model. However, finding symmetries is generally extremely complicated, often being the result of insightful thinking. In this work, we complement human ingenuity with an algorithm. We leverage the classically efficient Clifford group to find symmetries for arbitrary many-body Hamiltonians via a graph representation. We demonstrate our method on random and physical Hamiltonians, with instances of up to one thousand qubits and demonstrate how our approach can provide deeper understanding of the model.