A Contextual-Aware Position Encoding for Sequential Recommendation

TL;DR

This paper introduces CAPE, a novel position encoding method tailored for sequential recommendation models, significantly improving their performance and effectiveness in both benchmark datasets and real-world industrial applications.

Contribution

CAPE is the first position encoding specifically designed for sequential recommendation, enhancing model performance across various backbone architectures.

Findings

CAPE consistently improves multiple SR models on benchmark datasets.

CAPE achieves state-of-the-art results in experiments.

CAPE proves effective in real-world industrial deployment.

Abstract

Sequential recommendation (SR), which encodes user activity to predict the next action, has emerged as a widely adopted strategy in developing commercial personalized recommendation systems. A critical component of modern SR models is the attention mechanism, which synthesizes users' historical activities. This mechanism is typically order-invariant and generally relies on position encoding (PE). Conventional SR models simply assign a learnable vector to each position, resulting in only modest gains compared to traditional recommendation models. Moreover, limited research has been conducted on position encoding tailored for sequential recommendation, leaving a significant gap in addressing its unique requirements. To bridge this gap, we propose a novel Contextual-Aware Position Encoding method for sequential recommendation, abbreviated as CAPE. To the best of our knowledge, CAPE is the first PE method specifically designed for sequential recommendation. Comprehensive experiments conducted on benchmark SR datasets demonstrate that CAPE consistently enhances multiple mainstream backbone models and achieves state-of-the-art performance, across small and large scale model size. Furthermore, we deployed CAPE in an industrial setting on a real-world commercial platform, clearly showcasing the effectiveness of our approach. Our source code is available at https://github.com/yjdy/CAPE.