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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Recently, user-oriented auto-encoders (UAEs) have been widely used inrecommender systems to learn semantic representations of users based on theirhistorical ratings. However, since latent item variables are not modeled inUAE, it is difficult to utilize the widely available item content informationwhen ratings are sparse. In addition, whenever new items arrive, we need towait for collecting rating data for these items and retrain the UAE fromscratch, which is inefficient in practice. Aiming to address the above twoproblems simultaneously, we propose a mutually-regularized dual collaborativevariational auto-encoder (MD-CVAE) for recommendation. First, by replacingrandomly initialized last layer weights of the vanilla UAE with stacked latentitem embeddings, MD-CVAE integrates two heterogeneous information sources,i.e., item content and user ratings, into the same principled variationalframework where the weights of UAE are regularized by item content such thatconvergence to a non-optima due to data sparsity can be avoided. In addition,the regularization is mutual in that user ratings can also help the dual itemcontent module learn more recommendation-oriented item content embeddings.Finally, we propose a symmetric inference strategy for MD-CVAE where the firstlayer weights of the UAE encoder are tied to the latent item embeddings of theUAE decoder. Through this strategy, no retraining is required to recommendnewly introduced items. Empirical studies show the effectiveness of MD-CVAE inboth normal and cold-start scenarios. Codes are available athttps://github.com/yaochenzhu/MD-CVAE.

Mutually-Regularized Dual Collaborative Variational Auto-encoder for Recommendation Systems