Revisiting Collaborative Filtering

Hao Zhu, Yifei Zhang, Irwin King, Piotr Koniusz

Keywords: Collaborative Filtering

TL;DR: The design of loss functions for collaborative filtering can be done

Abstract: As a critical role in recommender systems, Collaborative Filtering (CF) is an indispensable technique. Its learning process typically comprises two key components: interaction encoder and loss function. Existing studies mainly concentrate on developing more robust encoders, such as graph neural networks, to acquire representations modelling multi-hop connectivity from user-item interactions. However, limited attention has been given to examining the desired characteristics of representations in collaborative filtering (CF).

Primary Area: representation learning for computer vision, audio, language, and other modalities

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 5249