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

Time series widely exists in real-world applications and many deep learningmodels have performed well on it. Current research has shown the importance oflearning strategy for models, suggesting that the benefit is the order and sizeof learning samples. However, no effective strategy has been proposed for timeseries due to its abstract and dynamic construction. Meanwhile, the existingone-shot tasks and continuous tasks for time series necessitate distinctlearning processes and mechanisms. No all-purpose approach has been suggested.In this work, we propose a novel Curricular and CyclicaL loss (CRUCIAL) tolearn time series for the first time. It is model- and task-agnostic and can beplugged on top of the original loss with no extra procedure. CRUCIAL has twocharacteristics: It can arrange an easy-to-hard learning order by dynamicallydetermining the sample contribution and modulating the loss amplitude; It canmanage a cyclically changed dataset and achieve an adaptive cycle bycorrelating the loss distribution and the selection probability. We prove thatcompared with monotonous size, cyclical size can reduce expected error.Experiments on 3 kinds of tasks and 5 real-world datasets show the benefits ofCRUCIAL for most deep learning models when learning time series.

Curricular and Cyclical Loss for Time Series Learning Strategy