arXiv subjects

Yuichiro Minato

Publications and source records attributed to Yuichiro Minato.

13 recordsLinked to original sources

arXiv2026

Hardware-Efficient Exchange-Only QML: Singlet-Triplet Spin Chains via Inter-pair Coupling without Magnetic Gradients

Standard universal quantum computing using exchange-only qubits typically requires three physical spins per logical qubit, leading to significant hardware overhead. Conversely, two-spin units offer higher density but rely on local magnetic field gradients for control, increasing integration complexity. In this paper, we propose a resource-efficient quantum machine learning (QML) architecture that achieves high expressibility using minimal two-spin units and Heisenberg exchange interactions alone, without any magnetic gradients. We shift the paradigm from universal gate-based control to utilizing the intrinsic, time-domain dynamics of a spin chain as a learning resource. Numerical simulations on MNIST digit classification demonstrate that the symmetry-protected constraints of isolated spin pairs are bypassed by leveraging inter-pair exchange coupling. This interference-mediated state mixing significantly enhances the expressibility of the Hilbert space. The model reaches a test-set accuracy of 90.9% +/- 0.2% over five independent seeds on the full 10,000-image MNIST test set. Under an identical linear readout, the trained quantum feature map (88.1%) clearly outperforms a classical linear baseline on the same PCA inputs (83.2%) as well as an untrained (reservoir-style) version of the same dynamics (53.0%), demonstrating that the learned, input-dependent exchange pulses implement a genuinely non-linear and trainable feature map. The protocol is also robust to experimentally relevant imperfections: accuracy remains at 89.9% under 10% quasi-static pulse-area noise and at 89.8% when every observable is estimated from 10^3 measurement shots. These findings suggest that competitive QML can be executed on the simplest possible semiconductor spin-chain hardware, bypassing the need for leakage-prone encodings or complex micro-magnet integration.

quant-ph↗

arXiv2025

Quantum Optimization-Based Route Compression for Efficient Navigation Systems

We present a novel quantum optimization-based route compression technique that significantly reduces storage requirements compared to conventional methods. Route optimization systems face critical challenges in efficiently storing selected routes, particularly under memory constraints. Our proposed method enhances route information compression rates by leveraging Higher Order Binary Optimization (HOBO), an extended formulation of Quadratic Unconstrained Binary Optimization (QUBO) commonly employed in quantum approximate optimization algorithms (QAOA) for combinatorial optimization problems. We implemented HOBO on real world map data and conducted comparative analysis between the traditional Ramer-Douglas-Peucker (RDP) algorithm and our proposed method. Results demonstrate that our approach successfully identifies yielding improved compression efficiency that scales with data size from candidate routes. Experimental validation confirms the technique viability for practical applications in navigation systems where memory constraints are critical. The HOBO formulation allows for representation of complex route that would be difficult to capture using classical compression algorithms. Our implementation demonstrates up to 30% improvement in compression rates while maintaining route fidelity within acceptable navigation parameters. This approach opens new possibilities for implementing quantum inspired optimization in transportation systems, potentially providing more efficient navigation services. This work represents a significant advancement in applying quantum optimization principles to practical transportation challenges.

quant-ph↗

arXiv2025

A Simplification Method for Inequality Constraints in Integer Binary Encoding HOBO Formulations

This study proposes a novel method for simplifying inequality constraints in Higher-Order Binary Optimization (HOBO) formulations. The proposed method addresses challenges associated with Quadratic Unconstrained Binary Optimization (QUBO) formulations, specifically the increased computational complexity and reduced solution accuracy caused by the introduction of slack variables and the resulting growth in auxiliary qubits. By efficiently integrating constraints, the method enhances the computational efficiency and accuracy of both quantum and classical solvers. The effectiveness of the proposed approach is demonstrated through numerical experiments applied to combinatorial optimization problems. The results indicate that this method expands the applicability of quantum algorithms to high-dimensional problems and improves the practicality of classical optimization solvers for optimization problems involving inequality constraints.

math.OC↗

arXiv2024

Numerical Exploration of the Pythagorean Theorem Using HOBO Algorithm

This paper introduces a novel method for finding integer sets that satisfy the Pythagorean theorem by leveraging the Higher-Order Binary Optimization (HOBO) formulation. Unlike the Quadratic Unconstrained Binary Optimization (QUBO) formulation, which struggles to express complex mathematical equations, HOBO's ability to model higher-order interactions between binary variables makes it well-suited for addressing more complex and expressive problem settings.

math.OC↗

arXiv2024

Two-Step QAOA: Enhancing Quantum Optimization by Decomposing K-hot Constraints in QUBO Formulations

The Quantum Approximate Optimization Algorithm (QAOA) has shown promise in solving combinatorial optimization problems by leveraging quantum computational power. We propose a simple approach, the Two-Step QAOA, which aims to improve the effectiveness of QAOA by decomposing problems with k-hot encoding QUBO (Quadratic Unconstrained Binary Optimization) formulations. By identifying and separating the problem into two stages, we transform soft constraints into hard constraints, simplifying the generation of initial conditions and enabling more efficient optimization. The method is particularly beneficial for tackling complex societal problems that often involve intricate constraint structures.

quant-ph↗

arXiv2024

HOBOTAN: Efficient Higher Order Binary Optimization Solver with Tensor Networks and PyTorch

In this study, we introduce HOBOTAN, a new solver designed for Higher Order Binary Optimization (HOBO). HOBOTAN supports both CPU and GPU, with the GPU version developed based on PyTorch, offering a fast and scalable system. This solver utilizes tensor networks to solve combinatorial optimization problems, employing a HOBO tensor that maps the problem and performs tensor contractions as needed. Additionally, by combining techniques such as batch processing for tensor optimization and binary-based integer encoding, we significantly enhance the efficiency of combinatorial optimization. In the future, the utilization of increased GPU numbers is expected to harness greater computational power, enabling efficient collaboration between multiple GPUs for high scalability. Moreover, HOBOTAN is designed within the framework of quantum computing, thus providing insights for future quantum computer applications. This paper details the design, implementation, performance evaluation, and scalability of HOBOTAN, demonstrating its effectiveness.

cs.MS↗

arXiv2024

Tensor Network Based HOBO Solver

In the field of quantum computing, combinatorial optimization problems are typically addressed using QUBO (Quadratic Unconstrained Binary Optimization) solvers. However, these solvers are often insufficient for tackling higher-order problems. In this paper, we introduce a novel and efficient solver designed specifically for HOBO (Higher-Order Binary Optimization) problem settings. Our approach leverages advanced techniques to effectively manage the complexity and computational demands associated with high-dimensional optimization tasks. The proposed solver is a promising tool with significant potential for future extensions in terms of formulation. This solver holds promising potential for a wide range of applications in quantum computing.

quant-ph↗

arXiv2024

Application of Tensorized Neural Networks for Cloud Classification

Convolutional neural networks (CNNs) have gained widespread usage across various fields such as weather forecasting, computer vision, autonomous driving, and medical image analysis due to its exceptional ability to extract spatial information, share parameters, and learn local features. However, the practical implementation and commercialization of CNNs in these domains are hindered by challenges related to model sizes, overfitting, and computational time. To address these limitations, our study proposes a groundbreaking approach that involves tensorizing the dense layers in the CNN to reduce model size and computational time. Additionally, we incorporate attention layers into the CNN and train it using Contrastive self-supervised learning to effectively classify cloud information, which is crucial for accurate weather forecasting. We elucidate the key characteristics of tensorized neural network (TNN), including the data compression rate, accuracy, and computational speed. The results indicate how TNN change their properties under the batch size setting.

cs.CV↗

arXiv2023

Workflow for practical quantum chemical calculations with quantum phase estimation algorithm: electronic ground and {\pi}-{\pi}* excited states of benzene and its derivatives{\dag}

Quantum computers are expected to perform the full-configuration interaction calculations with less computational resources compared to classical ones, thanks to the use of the quantum phase estimation (QPE) algorithms. However, only a limited number of the QPE-based quantum chemical calculations have been reported even for numerical simulations on a classical computer, and the practical workflow for the QPE computation has not yet been established. In this paper, we report the QPE simulations of the electronic ground and the {\pi}-{\pi}* excited singlet state of benzene and its chloro- and nitroderivatives as the representative industrially important systems, with the aid of GPGPU acceleration of quantum circuit simulations. We adopted the pseudo-natural orbitals obtained from the MP2 calculation as the basis for the wave function expansion, the CISD calculation within the active space to find the main electronic configurations to be included in the input wave function of the excited state, and the technique to reduce the truncation error the calculated total energies. The proposed computational workflow is easily applicable to other molecules and can be a standard approach for performing the QPE-based quantum chemical calculations of practical molecules.

quant-ph↗

arXiv2022

Explainable Natural Language Processing with Matrix Product States

Despite empirical successes of recurrent neural networks (RNNs) in natural language processing (NLP), theoretical understanding of RNNs is still limited due to intrinsically complex non-linear computations. We systematically analyze RNNs' behaviors in a ubiquitous NLP task, the sentiment analysis of movie reviews, via the mapping between a class of RNNs called recurrent arithmetic circuits (RACs) and a matrix product state (MPS). Using the von-Neumann entanglement entropy (EE) as a proxy for information propagation, we show that single-layer RACs possess a maximum information propagation capacity, reflected by the saturation of the EE. Enlarging the bond dimension beyond the EE saturation threshold does not increase model prediction accuracies, so a minimal model that best estimates the data statistics can be inferred. Although the saturated EE is smaller than the maximum EE allowed by the area law, our minimal model still achieves ~99% training accuracies in realistic sentiment analysis data sets. Thus, low EE is not a warrant against the adoption of single-layer RACs for NLP. Contrary to a common belief that long-range information propagation is the main source of RNNs' successes, we show that single-layer RACs harness high expressiveness from the subtle interplay between the information propagation and the word vector embeddings. Our work sheds light on the phenomenology of learning in RACs, and more generally on the explainability of RNNs for NLP, using tools from many-body quantum physics.

cond-mat.dis-nn↗

arXiv2020

Finding high-order Hadamard matrices by using quantum computers

Solving hard problems is one of the most important issues in computing to be addressed by a quantum computer. Previously, we have shown that the H-SEARCH; which is the problem of finding a Hadamard matrix (H-matrix) among all possible binary matrices of corresponding order, is a hard problem that can be solved by a quantum computer. However, due to the limitation on the number of qubits and connections in present day quantum processors, only low orders H-SEARCH are implementable. In this paper, we show that by adopting classical construction/search techniques of the H-matrix, we can develop new quantum computing methods to find higher order H-matrices. Especially, the Turyn-based quantum computing method can be further developed to find an arbitrarily high order H-matrix by balancing the classical and quantum resources. This method is potentially capable to find some unknown H-matrices of practical and scientific interests, where a classical computer alone cannot do because of the exponential grow of the complexity. We present some results of finding H-matrix of order more than one hundred and a prototypical experiment to find even higher order matrix by using the classical-quantum resource balancing method. Although heuristic optimizations generally only achieve approximate solutions, whereas the exact one should be determined by exhaustive listing; which is difficult to perform, in the H-SEARCH we can assure such exactness in polynomial time by checking the orthogonality of the solution. Since quantum advantage over the classical computing should have been measured by comparing the performance in solving a problem up to a definitive solution, the proposed method may lead to an alternate route for demonstrating practical quantum supremacy in the near future.

quant-ph↗

arXiv2019

Solving tiling puzzles with quantum annealing

To solve tiling puzzles, such as "pentomino" or "tetromino" puzzles, we need to find the correct solutions out of numerous combinations of rotations or piece locations. Solving this kind of combinatorial optimization problem is a very difficult problem in computational science, and quantum computing is expected to play an important role in this field. In this article, we propose a method and obtained specific formulas to find solutions for tetromino tiling puzzles using a quantum annealer. In addition, we evaluated these formulas using a simulator and using actual hardware DW2000Q.

quant-ph↗

arXiv2019

Finding Hadamard matrices by a quantum annealing machine

Finding a Hadamard matrix (H-matrix) among the set of all binary matrices of corresponding order is a hard problem, which potentially can be solved by quantum computing. We propose a method to formulate the Hamiltonian of finding H-matrix problem and address its implementation limitation on existing quantum annealing machine (QAM) that allows up to quadratic terms, whereas the problem naturally introduces higher order ones. For an M-order H-matrix, such a limitation increases the number of variables from M^2 to (M^3 +M^2-M)/2, which makes the formulation of the Hamiltonian too exhaustive to do by hand. We use symbolic computing techniques to manage this problem. Three related cases are discussed: (1) finding N < M orthogonal binary vectors, (2) finding M-orthogonal binary vectors, which is equivalent to finding a H-matrix, and (3) finding N-deleted vectors of an M-order H-matrix. Solutions of the problems by a 2-body simulated annealing software and by an actual quantum annealing hardware are also discussed.

quant-ph↗