Generative flow-based warm start of the variational quantum eigensolver

2025-07-02

AI Summary

VQE suffers from slow convergence on near-term quantum hardware due to complex objective landscapes and high optimization overhead. To address this, we propose Flow-VQE—a generative framework integrating conditional normalizing flows with parameterized quantum circuits, enabling gradient-free optimization and cross-system parameter transfer via preference learning. Flow-VQE is the first method to embed a generative flow model directly into the VQE optimization loop, providing a generalizable warm-start strategy. We validate it on molecular systems including H-chain, H₂O, NH₃, and C₆H₆. Compared to conventional VQE, Flow-VQE reduces circuit evaluations by one to two orders of magnitude; post-warm-start fine-tuning converges up to 50× faster. This significantly improves optimization efficiency and adaptability across diverse quantum chemical systems.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Hybrid quantum-classical algorithms like the variational quantum eigensolver (VQE) show promise for quantum simulations on near-term quantum devices, but are often limited by complex objective functions and expensive optimization procedures. Here, we propose Flow-VQE, a generative framework leveraging conditional normalizing flows with parameterized quantum circuits to efficiently generate high-quality variational parameters. By embedding a generative model into the VQE optimization loop through preference-based training, Flow-VQE enables quantum gradient-free optimization and offers a systematic approach for parameter transfer, accelerating convergence across related problems through warm-started optimization. We compare Flow-VQE to a number of standard benchmarks through numerical simulations on molecular systems, including hydrogen chains, water, ammonia, and benzene. We find that Flow-VQE outperforms baseline optimization algorithms, achieving computational accuracy with fewer circuit evaluations (improvements range from modest to more than two orders of magnitude) and, when used to warm-start the optimization of new systems, accelerates subsequent fine-tuning by up to 50-fold compared with Hartree--Fock initialization. Therefore, we believe Flow-VQE can become a pragmatic and versatile paradigm for leveraging generative modeling to reduce the costs of variational quantum algorithms.

Problem

• Improving convergence in variational quantum eigensolver optimization
• Reducing circuit evaluations for quantum simulations
• Enhancing parameter transfer across related quantum problems

Innovation

• Generative framework with conditional normalizing flows
• Quantum gradient-free optimization via preference training
• Parameter transfer for accelerated warm-start optimization