Recent developments in adaptive variational quantum algorithms

The performance of variational quantum algorithms relies on the variational ansatz, which generally needs to be specific to the target problem for easier optimization, and at the same time feasible to be run on noisy quantum processors. To efficiently incorporate the knowledge about the problem into the ansatz, an Adaptive Derivative-Assembled Problem-Tailored (ADAPT) strategy has been developed. It constructs the ansatz iteratively based on the gradient of the objective function, producing compact ansatze and saving resources. In this talk I will briefly review the application of the ADAPT strategy in both variational quantum eigensolver (VQE) and quantum approximation optimization algorithm (QAOA). Then I will discuss some recent developments in terms of operator selection and the impact of the initial state.