Trade-off between Gradient Measurement Efficiency and Expressivity in Deep Quantum Neural Networks
arXiv:2406.18316 · doi:10.1038/s41534-025-01036-7
Abstract
Quantum neural networks (QNNs) require an efficient training algorithm to achieve practical quantum advantages. A promising approach is gradient-based optimization, where gradients are estimated by quantum measurements. However, QNNs currently lack general quantum algorithms for efficiently measuring gradients, which limits their scalability. To elucidate the fundamental limits and potentials of efficient gradient estimation, we rigorously prove a trade-off between gradient measurement efficiency (the mean number of simultaneously measurable gradient components) and expressivity in deep QNNs. This trade-off indicates that more expressive QNNs require higher measurement costs per parameter for gradient estimation, while reducing QNN expressivity to suit a given task can increase gradient measurement efficiency. We further propose a general QNN ansatz called the stabilizer-logical product ansatz (SLPA), which achieves the trade-off upper bound by exploiting the symmetric structure of the quantum circuit. Numerical experiments show that the SLPA drastically reduces the sample complexity needed for training while maintaining accuracy and trainability compared to well-designed circuits based on the parameter-shift method.
32 pages, 11 figures
References in corpus (46)
- Fault-tolerant quantum computation by anyons
- Quantum Machine Learning
- A variational eigenvalue solver on a quantum processor
- Variational Quantum Algorithms
- Barren plateaus in quantum neural network training landscapes
- Solving the Quantum Many-Body Problem with Artificial Neural Networks
- E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
- A Quantum Approximate Optimization Algorithm
- Quantum Circuit Learning
- An introduction to quantum machine learning
- Parameterized quantum circuits as machine learning models
- Evaluating analytic gradients on quantum hardware
- Circuit-centric quantum classifiers
- Expressibility and entangling capability of parameterized quantum circuits for hybrid quantum-classical algorithms
- Exact and Approximate Unitary 2-Designs: Constructions and Applications
- Surface code quantum computing by lattice surgery
- Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
- Evenly distributed unitaries: on the structure of unitary designs
- Quantum Low-Density Parity-Check Codes
- Qulacs: a fast and versatile quantum circuit simulator for research purpose
- Differentiable Learning of Quantum Circuit Born Machine
- Diagnosing Barren Plateaus with Tools from Quantum Optimal Control
- Exploiting symmetry in variational quantum machine learning
- Theory of overparametrization in quantum neural networks
- Group-Invariant Quantum Machine Learning
- A Lie Algebraic Theory of Barren Plateaus for Deep Parameterized Quantum Circuits
- General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian
- Pseudorandom States, Non-Cloning Theorems and Quantum Money
- Optimizing quantum optimization algorithms via faster quantum gradient computation
- Theory for Equivariant Quantum Neural Networks
- Theoretical Guarantees for Permutation-Equivariant Quantum Neural Networks
- Symmetry Principles in Quantum Systems Theory
- On the Universality of the Quantum Approximate Optimization Algorithm
- Quantum approximate optimization is computationally universal
- Restrictions on realizable unitary operations imposed by symmetry and locality
- The Adjoint Is All You Need: Characterizing Barren Plateaus in Quantum Ansätze
- Quantum Graph Neural Networks
- Building spatial symmetries into parameterized quantum circuits for faster training
- Speeding up Learning Quantum States through Group Equivariant Convolutional Quantum Ansätze
- Efficient Simulation of Random Quantum States and Operators
- Representation Theory for Geometric Quantum Machine Learning
- On quantum backpropagation, information reuse, and cheating measurement collapse
- Splitting and Parallelizing of Quantum Convolutional Neural Networks for Learning Translationally Symmetric Data
- Quantum Deep Learning
- Post-variational quantum neural networks
- Training-efficient density quantum machine learning