GLINT-RU: Gated Lightweight Intelligent Recurrent Units for Sequential Recommender Systems

TL;DR

GLINT-RU introduces a lightweight recurrent unit with a dense gating mechanism and positional feature integration, significantly improving efficiency and accuracy in sequential recommender systems.

Contribution

It proposes a novel Gated Recurrent Unit with dense selective gating and a parallel positional feature infusion, enhancing both efficiency and recommendation quality.

Findings

Outperforms RNN, Transformer, MLP, and SSM baselines in accuracy

Achieves faster inference with comparable or better performance

Effective in capturing temporal and positional information

Abstract

Transformer-based models have gained significant traction in sequential recommender systems (SRSs) for their ability to capture user-item interactions effectively. However, these models often suffer from high computational costs and slow inference. Meanwhile, existing efficient SRS approaches struggle to embed high-quality semantic and positional information into latent representations. To tackle these challenges, this paper introduces GLINT-RU, a lightweight and efficient SRS leveraging a single-layer dense selective Gated Recurrent Units (GRU) module to accelerate inference. By incorporating a dense selective gate, GLINT-RU adaptively captures temporal dependencies and fine-grained positional information, generating high-quality latent representations. Additionally, a parallel mixing block infuses fine-grained positional features into user-item interactions, enhancing both recommendation quality and efficiency. Extensive experiments on three datasets demonstrate that GLINT-RU achieves superior prediction accuracy and inference speed, outperforming baselines based on RNNs, Transformers, MLPs, and SSMs. These results establish GLINT-RU as a powerful and efficient solution for SRSs.