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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Sequential recommendation (SR) investigates the dynamic user preferencesmodeling and generates the next-item prediction. The next item preference istypically generated by the affinity between the sequence and itemrepresentations. However, both sequence and item representations suffer fromthe rank degeneration issue due to the data sparsity problem. The rankdegeneration issue significantly impairs the representations for SR. Thismotivates us to measure how severe is the rank degeneration issue and alleviatethe sequence and item representation rank degeneration issues simultaneouslyfor SR.  In this work, we theoretically connect the sequence representationdegeneration issue with the item rank degeneration, particularly for shortsequences and cold items. We also identify the connection between the fastsingular value decay phenomenon and the rank collapse issue in transformersequence output and item embeddings. We propose the area under the singularvalue curve metric to evaluate the severity of the singular value decayphenomenon and use it as an indicator of rank degeneration. We furtherintroduce a novel singular spectrum smoothing regularization to alleviate therank degeneration on both sequence and item sides, which is the SingularsPectrum sMoothing for sequential Recommendation (SPMRec). We also establish acorrelation between the ranks of sequence and item embeddings and the rank ofthe user-item preference prediction matrix, which can affect recommendationdiversity. We conduct experiments on four benchmark datasets to demonstrate thesuperiority of SPMRec over the state-of-the-art recommendation methods,especially in short sequences. The experiments also demonstrate a strongconnection between our proposed singular spectrum smoothing and recommendationdiversity.

Addressing the Rank Degeneration in Sequential Recommendation via Singular Spectrum Smoothing