Jian Sun

TL;DR: Zhang et al. as mentioned in this paper proposed a new setup of differentiable architecture search (DARTS) with only training batchNorm, which dilutes the auxiliary-iary connection role of skip-connection in supernet optimization and enable search algorithm focus on fairer operation.

TL;DR: Zhang et al. as discussed by the authors proposed a Domain Projection and Contrastive Learning (DPCL) approach for generalized semantic segmentation, which includes two modules: self-supervised source domain projection (SSDP) and multi-level contrastive learning (MLCL) to reduce domain gap by projecting data to the source domain, while MLCL is a learning scheme to learn discriminative and generalizable features on the projected data.

TL;DR: In this article, a new concept of knowledge decomposition is introduced to further improve the performance of knowledge distillation, and the authors further put forward the \textbf{P}artial to Whole \textBF{Knowledge \text BF{D}istillation~(PWKD}) paradigm, which reconstructs teacher into weight-sharing sub-networks with same depth but increasing channel width, and train subnetworks jointly to obtain decomposed knowledge.

TL;DR: In this article, the authors introduced the first Mobile AI challenge, where the target is to develop an end-to-end deep learning-based image denoising solution that can demonstrate high efficiency on smartphone GPUs.