Zhilu Zhang

TL;DR: In this paper, a theoretically grounded set of noise-robust loss functions that can be seen as a generalization of mean absolute error (MAE) and categorical cross entropy (CCE) loss is proposed.

TL;DR: In this paper, a theoretically grounded set of noise-robust loss functions that can be seen as a generalization of mean absolute error (MAE) and categorical cross entropy (CCE) loss is proposed.

TL;DR: In this article, an instance-specific label smoothing technique was proposed to promote predictive diversity without the need for a separately trained teacher model. And they provided an empirical evaluation of the proposed method.

TL;DR: This paper uses the SVHN, CIFar-10 and CIFAR-100 datasets to empirically compare model diversity and confidence errors obtained using various dropout techniques, and shows the merit of structured dropout in a Bayesian active learning application.

TL;DR: This paper reviews the NTIRE 2022 Challenge on Super-Resolution and Quality Enhancement of Compressed Video and proposed the LDV 2.0 dataset, which includes theLDV dataset (240 videos) and 95 additional videos.