Accelerating Generative Recommendation via Simple Categorical User Sequence Compression

TL;DR

This paper introduces a straightforward method for compressing user history sequences in generative recommenders using item categories, significantly reducing computational costs while maintaining or improving accuracy.

Contribution

The proposed categorical compression technique efficiently reduces sequence length and computational load without sacrificing recommendation quality, outperforming existing models like HSTU.

Findings

Up to 6x reduction in computational cost

Up to 39% increase in accuracy at similar sequence length

Effective on large-scale datasets

Abstract

Although generative recommenders demonstrate improved performance with longer sequences, their real-time deployment is hindered by substantial computational costs. To address this challenge, we propose a simple yet effective method for compressing long-term user histories by leveraging inherent item categorical features, thereby preserving user interests while enhancing efficiency. Experiments on two large-scale datasets demonstrate that, compared to the influential HSTU model, our approach achieves up to a 6x reduction in computational cost and up to 39% higher accuracy at comparable cost (i.e., similar sequence length).