Light: Science & Applications has recently published the latest research results of Associate Professor HU Run, a Chinese supervisor of ICARE. The paper titled "General Deep Learning Framework for Emissivity Engineering" presents the findings of Associate Prof. HU Run's team proposing a universal design framework based on reinforced deep learning. The framework can achieve material selection and structural parameter optimization of thermal radiators simultaneously, without prior knowledge of materials and structures.

The team used multi-layer thin film structures as an example to demonstrate the power of a universal design framework for efficient reverse design of thermal radiators across different application contexts. The thermal radiator designed using this framework achieved excellent performance in three different applications: thermal camouflage, radiation refrigeration, and gas detection. Moreover, the general framework is suitable for the design optimization of high-dimensional complex structures and has high scalability in the design parameters of thermal radiators. The proposed framework provides a reference for nonlinear optimal design problems, other than thermal metamaterials, and promotes the practical application of emissivity engineering in thermal management and energy utilization.

Associate Prof. HU Run and his team have dedicated themselves to research in the fields of heat and mass transfer, thermal super-structural materials, functional devices, and thermal management of optoelectronic devices for several years. Recently, he has published numerous papers in reputed international journals such as Nature Communications, Science Advances, Physical Review X, Light: Science & Applications, Advanced Materials, Advanced Energy Materials, Materials Today, Nano Energy, and more.

This work was supported by the National Natural Science Foundation of China (NSFC) grants 52211540005 and 52076087 as well as other funding sources.