RaSeRec: Retrieval-Augmented Sequential Recommendation

TL;DR

RaSeRec introduces a retrieval-augmented framework for sequential recommendation that maintains a dynamic memory bank to better adapt to evolving user preferences and recall long-tail patterns, outperforming existing models.

Contribution

This work proposes RaSeRec, a novel retrieval-augmented SeRec framework with a dynamic memory bank and two-stage training, addressing preference drift and implicit memory limitations.

Findings

RaSeRec outperforms baseline models on three datasets.

The dynamic memory bank effectively captures preference changes.

Retrieval augmentation improves long-tail pattern recall.

Abstract

Although prevailing supervised and self-supervised learning augmented sequential recommendation (SeRec) models have achieved improved performance with powerful neural network architectures, we argue that they still suffer from two limitations: (1) Preference Drift, where models trained on past data can hardly accommodate evolving user preference; and (2) Implicit Memory, where head patterns dominate parametric learning, making it harder to recall long tails. In this work, we explore retrieval augmentation in SeRec, to address these limitations. Specifically, we propose a Retrieval-Augmented Sequential Recommendation framework, named RaSeRec, the main idea of which is to maintain a dynamic memory bank to accommodate preference drifts and retrieve relevant memories to augment user modeling explicitly. It consists of two stages: (i) collaborative-based pre-training, which learns to recommend and retrieve; (ii) retrieval-augmented fine-tuning, which learns to leverage retrieved memories. Extensive experiments on three datasets fully demonstrate the superiority and effectiveness of RaSeRec. The implementation code is available at https://github.com/HITsz-TMG/RaSeRec.