English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Data uncertainty is commonly observed in the images for face recognition(FR). However, deep learning algorithms often make predictions with highconfidence even for uncertain or irrelevant inputs. Intuitively, FR algorithmscan benefit from both the estimation of uncertainty and the detection ofout-of-distribution (OOD) samples. Taking a probabilistic view of the currentclassification model, the temperature scalar is exactly the scale ofuncertainty noise implicitly added in the softmax function. Meanwhile, theuncertainty of images in a dataset should follow a prior distribution. Based onthe observation, a unified framework for uncertainty modeling and FR, RandomTemperature Scaling (RTS), is proposed to learn a reliable FR algorithm. Thebenefits of RTS are two-fold. (1) In the training phase, it can adjust thelearning strength of clean and noisy samples for stability and accuracy. (2) Inthe test phase, it can provide a score of confidence to detect uncertain,low-quality and even OOD samples, without training on extra labels. Extensiveexperiments on FR benchmarks demonstrate that the magnitude of variance in RTS,which serves as an OOD detection metric, is closely related to the uncertaintyof the input image. RTS can achieve top performance on both the FR and OODdetection tasks. Moreover, the model trained with RTS can perform robustly ondatasets with noise. The proposed module is light-weight and only addsnegligible computation cost to the model.

Improving Training and Inference of Face Recognition Models via Random Temperature Scaling