Normalizing flows improve quantum circuit simulation accuracy at scale
Normalizing Flows for Classical Simulation of Quantum Circuits in Phase Space
Machine Learning
Summary
Simulating quantum circuits on regular computers is hard because quantum states get complicated quickly as they grow. This paper presents a new way to represent these quantum states as smooth probability distributions in a phase space, and then simulates quantum operations as transformations called normalizing flows between those distributions. The authors show that this approach can simulate several common types of quantum circuits more accurately than previous neural network-based methods, especially for large numbers of qubits. This could help developers better model quantum systems without needing a quantum computer.
What this means in practice
- •For quantum software developers: Improve the accuracy of classical simulators for testing and debugging quantum algorithms using flow-based models.
- •For machine learning engineers: Use phase space normalizing flows to model complex high-dimensional probability distributions for quantum-inspired generative tasks.
Authors
Edward Jiang, Timothy Heightman, Marcin Płodzień
Abstract
We present a novel framework to learn a representation of quantum circuits based on the phase space formulation of many-body systems. Given that $N$-qubit states can be represented as probability distributions on a smooth $2N$-dimensional manifold, we show how unitary gates can be parameterized as normalizing flows between these distributions, opening a new avenue for quantum circuit simulation using generative models. We evaluate our method on three common families of quantum circuits, and show that its accuracy exceeds that of current state-of-the-art neural network methods for large system sizes.