Towards Automatic Sampling of User Behaviors for Sequential Recommender Systems

2023-11-01

arXiv.org

AI Summary

Existing sequential recommendation systems treat user historical behaviors uniformly, ignoring their heterogeneous importance and thereby limiting recommendation accuracy. To address this, we propose AutoSAM—a novel framework that formulates behavioral importance modeling as an adaptive distribution learning problem, marking the first such approach in sequential recommendation. AutoSAM introduces a differentiable reinforcement learning sampling layer based on policy gradients, jointly optimizing future behavior prediction and sequence perplexity in an end-to-end manner. By dynamically identifying high-informativeness behavioral subsequences, it enhances model generalization. Extensive experiments on four real-world datasets against diverse state-of-the-art baselines demonstrate consistent improvements: average gains of 3.2%–5.8% in Recall@10 and MRR, significantly outperforming existing methods. The implementation code will be publicly released.

Abstract

Sequential recommender systems (SRS) have gained widespread popularity in recommendation due to their ability to effectively capture dynamic user preferences. One default setting in the current SRS is to uniformly consider each historical behavior as a positive interaction. Actually, this setting has the potential to yield sub-optimal performance, as each item makes a distinct contribution to the user's interest. For example, purchased items should be given more importance than clicked ones. Hence, we propose a general automatic sampling framework, named AutoSAM, to non-uniformly treat historical behaviors. Specifically, AutoSAM augments the standard sequential recommendation architecture with an additional sampler layer to adaptively learn the skew distribution of the raw input, and then sample informative sub-sets to build more generalizable SRS. To overcome the challenges of non-differentiable sampling actions and also introduce multiple decision factors for sampling, we further introduce a novel reinforcement learning based method to guide the training of the sampler. We theoretically design multi-objective sampling rewards including Future Prediction and Sequence Perplexity, and then optimize the whole framework in an end-to-end manner by combining the policy gradient. We conduct extensive experiments on benchmark recommender models and four real-world datasets. The experimental results demonstrate the effectiveness of the proposed approach. We will make our code publicly available after the acceptance.

Problem

• Sequential Recommendation
• User Behavior Importance
• Precision Improvement

Innovation

• AutoSAM
• Reinforcement Learning
• Personalized Recommendation