VQEzy: An Open-Source Dataset for Parameter Initialize in Variational Quantum Eigensolvers

2025-09-21

AI Summary

Variational Quantum Eigensolver (VQE) performance is highly sensitive to parameter initialization, yet existing machine learning approaches are hindered by small-scale, domain-specific, and incomplete datasets—typically comprising only hundreds of instances and lacking essential components such as Hamiltonians, ansatz circuits, and full optimization trajectories. Method: We introduce VQEzy, the first large-scale, open-source dataset for VQE initialization, spanning quantum chemistry, materials simulation, and quantum physics. It comprises 12,110 complete, standardized instances across seven task categories, each providing Hamiltonians, ansatz architectures, optimization algorithms, and full training trajectories. Contribution/Results: VQEzy establishes a high-quality, reproducible benchmark for VQE initialization, enabling cross-task generalization and rigorous evaluation. It significantly improves training efficiency and generalization capability of initialization strategies, serving as critical infrastructure for optimizing quantum algorithms in the NISQ era.

Technology Category

• Machine Learning: Quantum Machine Learning
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Variational Quantum Eigensolvers (VQEs) are a leading class of noisy intermediate-scale quantum (NISQ) algorithms, whose performance is highly sensitive to parameter initialization. Although recent machine learning-based initialization methods have achieved state-of-the-art performance, their progress has been limited by the lack of comprehensive datasets. Existing resources are typically restricted to a single domain, contain only a few hundred instances, and lack complete coverage of Hamiltonians, ansatz circuits, and optimization trajectories. To overcome these limitations, we introduce VQEzy, the first large-scale dataset for VQE parameter initialization. VQEzy spans three major domains and seven representative tasks, comprising 12,110 instances with full VQE specifications and complete optimization trajectories. The dataset is available online, and will be continuously refined and expanded to support future research in VQE optimization.

Problem

• Addressing limited datasets for VQE parameter initialization methods
• Overcoming incomplete coverage of Hamiltonians and ansatz circuits
• Providing a large-scale dataset with full optimization trajectories

Innovation

• Large-scale dataset for VQE parameter initialization
• Spans three domains and seven representative tasks
• Contains full VQE specifications and optimization trajectories