Pattern-wise Transparent Sequential Recommendation

2024-02-18

arXiv.org

AI Summary

To address the challenge of balancing model interpretability and recommendation performance in sequential recommendation, this paper proposes the Pattern-level Transparent Sequential Recommendation (PTSR) framework. PTSR unsupervisedly decomposes user behavioral sequences into multi-scale semantic patterns and models each pattern’s probabilistic contribution to the final recommendation. It employs a self-calibrating attention mechanism to implicitly learn pattern importance weights—without requiring auxiliary features or explicit annotations. As the first approach enabling end-to-end transparent decision-making at the pattern level, PTSR achieves both state-of-the-art recommendation accuracy and intrinsic interpretability. Extensive experiments on five public benchmarks demonstrate its superior performance over existing methods. Quantitative analysis and case studies further confirm its high accuracy in identifying critical behavioral patterns and provide clear, trustworthy decision rationales grounded in interpretable pattern activations.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Machine Learning: Transparent, Interpretable, Explainable ML
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Algorithmic accountability and transparency on the web

Abstract

A transparent decision-making process is essential for developing reliable and trustworthy recommender systems. For sequential recommendation, it means that the model can identify key items that account for its recommendation results. However, achieving both interpretability and recommendation performance simultaneously is challenging, especially for models that take the entire sequence of items as input without screening. In this paper, we propose an interpretable framework (named PTSR) that enables a pattern-wise transparent decision-making process without extra features. It breaks the sequence of items into multi-level patterns that serve as atomic units throughout the recommendation process. The contribution of each pattern to the outcome is quantified in the probability space. With a carefully designed score correction mechanism, the pattern contribution can be implicitly learned in the absence of ground-truth key patterns. The final recommended items are those that most key patterns strongly endorse. Extensive experiments on five public datasets demonstrate remarkable recommendation performance, while statistical analysis and case studies validate the model interpretability.

Problem

• Achieving interpretability in sequential recommendation systems
• Identifying key items influencing recommendation outcomes
• Balancing recommendation performance with model transparency

Innovation

• Pattern-wise transparent decision-making process
• Multi-level patterns as atomic units
• Score correction mechanism for pattern contribution