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

Soft dynamic time warping (SDTW) is a differentiable loss function thatallows for training neural networks from weakly aligned data. Typically, SDTWis used to iteratively compute and refine soft alignments that compensate fortemporal deviations between the training data and its weakly annotated targets.One major problem is that a mismatch between the estimated soft alignments andthe reference alignments in the early training stage leads to incorrectparameter updates, making the overall training procedure unstable. In thispaper, we investigate such stability issues by considering the task of pitchclass estimation from music recordings as an illustrative case study. Inparticular, we introduce and discuss three conceptually different strategies (ahyperparameter scheduling, a diagonal prior, and a sequence unfolding strategy)with the objective of stabilizing intermediate soft alignment results. Finally,we report on experiments that demonstrate the effectiveness of the strategiesand discuss efficiency and implementation issues.

Stabilizing Training with Soft Dynamic Time Warping: A Case Study for Pitch Class Estimation with Weakly Aligned Targets