Abstract:Negative sampling is essential for implicit-feedback-based collaborative filtering, which is used to constitute negative signals from massive unlabeled data to guide supervised learning. The state-of-the-art idea is to utilize hard negative samples that carry more useful information to form a better decision boundary. To balance efficiency and effectiveness, the vast majority of existing methods follow the two-pass approach, in which the first pass samples a fixed number of unobserved items by a simple static distribution and then the second pass selects the final negative items using a more sophisticated negative sampling strategy. However, selecting negative samples from the original items is inherently restricted, and thus may not be able to contrast positive samples well. In this paper, we confirm this observation via experiments and introduce two limitations of existing solutions: ambiguous trap and information discrimination. Our response to such limitations is to introduce augmented negative samples. This direction renders a substantial technical challenge because constructing unconstrained negative samples may introduce excessive noise that distorts the decision boundary. To this end, we introduce a novel generic augmented negative sampling paradigm and provide a concrete instantiation. First, we disentangle hard and easy factors of negative items. Next, we generate new candidate negative samples by augmenting only the easy factors in a regulated manner: the direction and magnitude of the augmentation are carefully calibrated. Finally, we design an advanced negative sampling strategy to identify the final augmented negative samples, which considers not only the score function used in existing methods but also a new metric called augmentation gain. Extensive experiments on real-world datasets demonstrate that our method significantly outperforms state-of-the-art baselines.

Augmented Negative Sampling for Collaborative Filtering

Simplify and Robustify Negative Sampling for Implicit Collaborative Filtering

Adaptive Hardness Negative Sampling for Collaborative Filtering

MixGCF

Optimizing Top-N Collaborative Filtering Via Dynamic Negative Item Sampling

Enhancing Recommender Systems: A Strategy to Mitigate False Negative Impact

Trading Hard Negatives and True Negatives: A Debiased Contrastive Collaborative Filtering Approach.

Ultimate Negative Sampling for Contrastive Learning

Bayesian Negative Sampling for Recommendation

Fairly Adaptive Negative Sampling for Recommendations.

Towards Automated Negative Sampling in Implicit Recommendation

Negative Sampling Strategy Based on Multi-Hop Neighbors for Graph Representation Learning

ABNS: Association-based negative sampling for collaborative filtering

A Gain-Tuning Dynamic Negative Sampler for Recommendation

Neighborhood-based Hard Negative Mining for Sequential Recommendation

Online Collaborative Filtering With Implicit Feedback

Rethinking Collaborative Metric Learning: Toward an Efficient Alternative without Negative Sampling

Revisiting Negative Sampling Vs. Non-sampling in Implicit Recommendation

AFANS: Augmentation-Free Graph Contrastive Learning with Adversarial Negative Sampling

Reinforcement negative sampling recommendation based on collaborative knowledge graph

Reinforced Negative Sampling for Recommendation with Exposure Data