Reproduce “Generative Contrastive Graph Learning for Recommendation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yonghui Yang, Zhengwei Wu, Le Wu, Kun Zhang, Richang Hong, Zhiqiang Zhang, Jun Zhou, Meng Wang
VGCL addresses sparse recommendation data by learning a Gaussian distribution for each user and item from the interaction graph, sampling alternate representations, and encouraging agreement within nodes and clusters. On three public datasets, it outscored comparison methods, including SimGCL; on Douban-Book, its top-20 ranking score was 0.1638 versus 0.1540. These are offline results from random splits, not deployed recommendations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.