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

We propose a simple but efficient method termed Guided Learning forweakly-labeled semi-supervised sound event detection (SED). There are twosub-targets implied in weakly-labeled SED: audio tagging and boundarydetection. Instead of designing a single model by considering a trade-offbetween the two sub-targets, we design a teacher model aiming at audio taggingto guide a student model aiming at boundary detection to learn using theunlabeled data. The guidance is guaranteed by the audio tagging performance gapof the two models. In the meantime, the student model liberated from thetrade-off is able to provide more excellent boundary detection results. Wepropose a principle to design such two models based on the relation between thetemporal compression scale and the two sub-targets. We also propose anend-to-end semi-supervised learning process for these two models to enabletheir abilities to rise alternately. Experiments on the DCASE2018 Task4 datasetshow that our approach achieves competitive performance.

Guided learning for weakly-labeled semi-supervised sound event detection