Automated Quantum Algorithm Design via Evolutionary Search and Hierarchical Quantum Circuit Representation

Quantum algorithms are inherently modular and often exhibit repeating patterns. We exploit this by utilizing a hierarchical quantum circuit representation [1] in conjunction with evolutionary search for the automated design of such algorithms. In this representation, quantum circuits are abstracted beyond the usual gate-sequence description and scale automatically to any circuit size. This allows us to evaluate a single candidate algorithm on varying problem sizes, which enables global features such as the optimal number of repetitions and parameter relationships to be learnt. We present our method as a general tool that uses techniques from Neural Architecture Search and discuss its performance compared to an exhaustive search. Remarkably, we are able to rediscover three well known quantum algorithms, the Quantum Fourier Transform, Deutsch-Jozsa and Grover's search.

[1] Lourens, M., Sinayskiy, I., Park, D. K., Blank, C., & Petruccione, F. (2023). Hierarchical quantum circuit representations for neural architecture search. npj Quantum Information, 9(1), 79