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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Beyond assigning the correct class, an activity recognition model should alsobe able to determine, how certain it is in its predictions. We present thefirst study of how welthe confidence values of modern action recognitionarchitectures indeed reflect the probability of the correct outcome and proposea learning-based approach for improving it. First, we extend two popular actionrecognition datasets with a reliability benchmark in form of the expectedcalibration error and reliability diagrams. Since our evaluation highlightsthat confidence values of standard action recognition architectures do notrepresent the uncertainty well, we introduce a new approach which learns totransform the model output into realistic confidence estimates through anadditional calibration network. The main idea of our Calibrated ActionRecognition with Input Guidance (CARING) model is to learn an optimal scalingparameter depending on the video representation. We compare our model with thenative action recognition networks and the temperature scaling approach - awide spread calibration method utilized in image classification. Whiletemperature scaling alone drastically improves the reliability of theconfidence values, our CARING method consistently leads to the best uncertaintyestimates in all benchmark settings.

Uncertainty-sensitive Activity Recognition: a Reliability Benchmark and the CARING Models