Retrieving GNN Architecture for Collaborative Filtering

By F. Liang et al

Published on June 10, 2023

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Related Works
3. Methodology
3.1 Problem Definition
3.2 Neural Retrieval Approach
4. Experimental Results
5. Conclusion

Summary

The document discusses the problem of rapidly obtaining a well-performing GNN architecture for new recommendation scenarios. It introduces a novel neural retrieval approach, RGCF, based on meta-learning and retrieval paradigm. The approach uses two-level meta-features, ranking loss, and task-level data augmentation to search for architectures efficiently. Experimental results show RGCF outperforms manual and NAS baselines in rating prediction and item ranking tasks, significantly improving search efficiency.