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

Self-supervised representation learning can mitigate the limitations inrecognition tasks with few manually labeled data but abundant unlabeleddata---a common scenario in sound event research. In this work, we exploreunsupervised contrastive learning as a way to learn sound eventrepresentations. To this end, we propose to use the pretext task of contrastingdifferently augmented views of sound events. The views are computed primarilyvia mixing of training examples with unrelated backgrounds, followed by otherdata augmentations. We analyze the main components of our method via ablationexperiments. We evaluate the learned representations using linear evaluation,and in two in-domain downstream sound event classification tasks, namely, usinglimited manually labeled data, and using noisy labeled data. Our resultssuggest that unsupervised contrastive pre-training can mitigate the impact ofdata scarcity and increase robustness against noisy labels, outperformingsupervised baselines.

Unsupervised Contrastive Learning of Sound Event Representations