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

Training good representations for items is critical in recommender models.Typically, an item is assigned a unique randomly generated ID, and is commonlyrepresented by learning an embedding corresponding to the value of the randomID. Although widely used, this approach have limitations when the number ofitems are large and items are power-law distributed -- typical characteristicsof real-world recommendation systems. This leads to the item cold-startproblem, where the model is unable to make reliable inferences for tail andpreviously unseen items. Removing these ID features and their learnedembeddings altogether to combat cold-start issue severely degrades therecommendation quality. Content-based item embeddings are more reliable, butthey are expensive to store and use, particularly for users' past iteminteraction sequence. In this paper, we use Semantic IDs, a compact discreteitem representations learned from content embeddings using RQ-VAE that captureshierarchy of concepts in items. We showcase how we use them as a replacement ofitem IDs in a resource-constrained ranking model used in an industrial-scalevideo sharing platform. Moreover, we show how Semantic IDs improves thegeneralization ability of our system, without sacrificing top-level metrics.

Better Generalization with Semantic IDs: A case study in Ranking for Recommendations