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

A common challenge in personalized user preference prediction is thecold-start problem. Due to the lack of user-item interactions, directlylearning from the new users' log data causes serious over-fitting problem.Recently, many existing studies regard the cold-start personalized preferenceprediction as a few-shot learning problem, where each user is the task andrecommended items are the classes, and the gradient-based meta learning method(MAML) is leveraged to address this challenge. However, in real-worldapplication, the users are not uniformly distributed (i.e., different users mayhave different browsing history, recommended items, and user profiles. Wedefine the major users as the users in the groups with large numbers of userssharing similar user information, and other users are the minor users),existing MAML approaches tend to fit the major users and ignore the minorusers. To address this cold-start task-overfitting problem, we propose a novelpersonalized adaptive meta learning approach to consider both the major and theminor users with three key contributions: 1) We are the first to present apersonalized adaptive learning rate meta-learning approach to improve theperformance of MAML by focusing on both the major and minor users. 2) Toprovide better personalized learning rates for each user, we introduce asimilarity-based method to find similar users as a reference and a tree-basedmethod to store users' features for fast search. 3) To reduce the memory usage,we design a memory agnostic regularizer to further reduce the space complexityto constant while maintain the performance. Experiments on MovieLens,BookCrossing, and real-world production datasets reveal that our methodoutperforms the state-of-the-art methods dramatically for both the minor andmajor users.

Personalized Adaptive Meta Learning for Cold-start User Preference Prediction