My major research interest is to improve the effectiveness and efficiency of deep neural networks via knowledge distillation. In 2019, I firstly propose the idea of self-distillation. So far, I have also successfully applied knowledge distillation to all kinds of tasks, such as 2D classification detection, segmentation, image-to-image translation, video recognition, point cloud-based detection, BEV based multi-view detection, representation and so on. Please refer to my CV for more information.