Special-Unitary Parameterization for Trainable Variational Quantum Circuits

2025-07-07

AI Summary

To address the barren-plateau problem—exponential gradient vanishing—in hardware-efficient variational quantum circuits (VQCs), this work proposes SUN-VQC: a symmetry-constrained architecture that restricts quantum evolution to symmetric subspaces of the SU(2ᵏ) subgroup, thereby fundamentally reducing the Lie algebra dimension and mitigating exponential gradient decay. Instead of conventional ansätze prone to barren plateaus, SUN-VQC employs Lie subalgebra engineering to construct trainable circuits. It leverages single-exponential SU(2ᵏ) layers and generalized parameter-shift rules for unbiased, ancilla-free, hardware-compatible gradient estimation. Experiments on quantum autoencoding and classification tasks demonstrate a tenfold increase in gradient magnitude, 2–3× faster convergence, and higher final fidelity. The architecture is scalable and underpinned by rigorous theoretical analysis.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

We propose SUN-VQC, a variational-circuit architecture whose elementary layers are single exponentials of a symmetry-restricted Lie subgroup, $mathrm{SU}(2^{k}) subset mathrm{SU}(2^{n})$ with $k ll n$. Confining the evolution to this compact subspace reduces the dynamical Lie-algebra dimension from $mathcal{O}(4^{n})$ to $mathcal{O}(4^{k})$, ensuring only polynomial suppression of gradient variance and circumventing barren plateaus that plague hardware-efficient ansätze. Exact, hardware-compatible gradients are obtained using a generalized parameter-shift rule, avoiding ancillary qubits and finite-difference bias. Numerical experiments on quantum auto-encoding and classification show that SUN-VQCs sustain order-of-magnitude larger gradient signals, converge 2--3$ imes$ faster, and reach higher final fidelities than depth-matched Pauli-rotation or hardware-efficient circuits. These results demonstrate that Lie-subalgebra engineering provides a principled, scalable route to barren-plateau-resilient VQAs compatible with near-term quantum processors.

Problem

• Reduces gradient variance in variational quantum circuits
• Avoids barren plateaus in hardware-efficient ansatze
• Enables scalable, plateau-resilient quantum algorithms

Innovation

• Uses SU(2^k) subgroup for compact parameterization
• Implements generalized parameter-shift rule
• Achieves faster convergence and higher fidelity