Orqviz: Visualizing High-Dimensional Landscapes in Variational Quantum Algorithms

Variational Quantum Algorithms (VQAs) are promising candidates for finding practical applications of near to mid-term quantum computers. To best utilize available quantum resources, it is vitally important that we do not treat VQAs as black boxes. Towards this end, there has been an increasing effort to study the intricacies of VQAs, including the presence or absence of barren plateaus, the expressivity of circuit ansätze, and heuristics for parameter initialization. Many of these studies can be linked to the properties of the optimization loss landscape explored for each algorithm. In our work, we present a variety of techniques for visualizing and analyzing the high-dimensional loss landscapes of VQAs. We review and apply the techniques to three examples with diverse application domains: the Quantum Approximate Optimization Algorithm, the Quantum Circuit Born Machine, and the Variational Quantum Eigensolver. Additionally, we include an investigation on the impact of noise due to finite sampling in the estimation of the loss functions. This work is accompanied by the release of the open-source Python package "orqviz", which supports all the techniques discussed. Orqviz enables flexible visual analysis of high-dimensional VQA landscapes.