Equivariant quantum circuits for learning on weighted graphs
This work addresses the scarcity of data-structure-motivated ansatzes in quantum machine learning, offering a domain-specific solution for graph-based learning tasks.
The authors tackled the problem of designing variational quantum circuits for machine learning on weighted graphs by introducing an ansatz that respects node permutation equivariance, and they demonstrated its effectiveness on neural combinatorial optimization tasks, with analytical and numerical results supporting improved performance.
Variational quantum algorithms are the leading candidate for advantage on near-term quantum hardware. When training a parametrized quantum circuit in this setting to solve a specific problem, the choice of ansatz is one of the most important factors that determines the trainability and performance of the algorithm. In quantum machine learning (QML), however, the literature on ansatzes that are motivated by the training data structure is scarce. In this work, we introduce an ansatz for learning tasks on weighted graphs that respects an important graph symmetry, namely equivariance under node permutations. We evaluate the performance of this ansatz on a complex learning task, namely neural combinatorial optimization, where a machine learning model is used to learn a heuristic for a combinatorial optimization problem. We analytically and numerically study the performance of our model, and our results strengthen the notion that symmetry-preserving ansatzes are a key to success in QML.