Abstract:The tremendous positive driving effect of Graph Convolutional Network (GCN) and Graph Contrastive Learning (GCL) for recommender systems has become a consensus. GCN encoders are extensively used in recommendation models for capturing high-order connectivities between users and items, whereas GCL accelerates the training of recommendation tasks by adding extra supervision signals from contrastive objectives. However, little attention has been paid on corresponding theories that are truly tailored to recommendation tasks. From the technical perspective, Collaborative Filtering (CF) is seen as an important factor in recommender systems. It is applied to measure user-user, item-item, and user-item similarities rather than to achieve better clustering or node classification results. Besides, heuristic-based data augmentation may not be hold true in the field of recommender systems as it requires additional training costs and introduces noises that will corrupt the interaction graph structure and the semantic information of nodes. To tackle these limitations, we propose a novel Embedding-less Graph Collaborative Filtering (EGCF) for recommendation, which is tailor-made for the problem mentioned for CF and further simplifies existing solutions. Structurally, it consists of two parts: embedding-less GCN and embedding-less GCL. The former improves user-item affinity by streamlining user-type embeddings and carrying out iterative graph convolution. And the latter utilizes three-type contrastive objectives to directly measure the alignment and the uniformity of users, items, and interaction pairs, respectively, avoiding any type of data augmentation or multi-view construction. Even though EGCF has been extremely streamlined, extensive experimental results on three classical datasets demonstrate the effectiveness of EGCF in terms of recommendation accuracy and training efficiency. The code and used datasets are released at https://github.com/BlueGhostYi/ID-GRec.

A General Strategy Graph Collaborative Filtering for Recommendation Unlearning

UltraRE: Enhancing RecEraser for Recommendation Unlearning Via Error Decomposition.

Selective and Collaborative Influence Function for Efficient Recommendation Unlearning

CURE4Rec: A Benchmark for Recommendation Unlearning with Deeper Influence

Federated Unlearning for On-Device Recommendation

On the Effectiveness of Unlearning in Session-Based Recommendation

Recommendation Unlearning via Influence Function

Making recommender systems forget: Learning and unlearning for erasable recommendation

Recommendation Unlearning via Matrix Correction

Towards Efficient and Effective Unlearning of Large Language Models for Recommendation

A Survey on Recommendation Unlearning: Fundamentals, Taxonomy, Evaluation, and Open Questions

Exact and Efficient Unlearning for Large Language Model-based Recommendation

Forgetting Fast in Recommender Systems

Community-Centric Graph Unlearning

Simplify to the Limit! Embedding-less Graph Collaborative Filtering for Recommender Systems

Unified Collaborative Filtering over Graph Embeddings

An Accelerator of Feature Selection Applying a General Fuzzy Rough Model.

RRL: Recommendation Reverse Learning

Efficient Heterogeneous Collaborative Filtering Without Negative Sampling for Recommendation.

Adaptive Clipping and Distillation Enabled Federated Unlearning