Machine learning for halide perovskite materials

Abstract

Halide perovskite materials serve as excellent candidates for solar cell and optoelectronic devices. Recently, the design of the halide perovskite materials is greatly facilitated by machine learning techniques, which effectively identify suitable halide perovskite candidates and unveil hidden relationships by algorithms that mimic the human cognitive functions. In this manuscript, we review recent progresses on the machine learning studies of the halide perovskite materials, including the prediction and understanding of lead-free and stable halide perovskite materials. The structural descriptors to describe the property and performance of the halide perovskite materials are discussed. In addition, the design strategy of the additive species for the halide perovskite materials via the machine learning technique is provided. Suggestions to further develop the halide perovskite-based systems via the machine learning methods in the future are provided.