RAIE: Region-Aware Incremental Preference Editing with LoRA for LLM-based Recommendation

2026-02-28

AI Summary

This work addresses the challenges of updating large language model–based recommender systems under dynamic user preferences, where conventional global fine-tuning or pointwise editing leads to imbalanced update granularity, catastrophic forgetting, and recommendation inconsistency. To this end, the authors propose a plug-and-play, region-aware incremental editing framework that introduces, for the first time, a region-level editing mechanism. Semantic-coherent user preference regions are constructed via spherical k-means clustering, and user sequences are routed to these regions through a confidence-gated mechanism. Each region is equipped with a dedicated LoRA adapter, enabling localized Update, Expand, and Add operations while keeping the backbone model frozen. Experiments on two benchmark datasets under temporal slicing protocols demonstrate that the proposed framework significantly outperforms existing methods, effectively mitigating catastrophic forgetting and enhancing both recommendation accuracy and stability.

Technology Category

• Data Mining & Knowledge Management: Recommender Systems
• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Large language models (LLMs) are increasingly adopted as the backbone of recommender systems. However, user-item interactions in real-world scenarios are non-stationary, making preference drift over time inevitable. Existing model update strategies mainly rely on global fine-tuning or pointwise editing, but they face two fundamental challenges: (i) imbalanced update granularity, where global updates perturb behaviors unrelated to the target while pointwise edits fail to capture broader preference shifts; (ii) unstable incremental updates, where repeated edits interfere with prior adaptations, leading to catastrophic forgetting and inconsistent recommendations. To address these issues, we propose Region-Aware Incremental Editing (RAIE), a plug-in framework that freezes the backbone model and performs region-level updates. RAIE first constructs semantically coherent preference regions via spherical k-means in the representation space. It then assigns incoming sequences to regions via confidence-aware gating and performs three localized edit operations - Update, Expand, and Add - to dynamically revise the affected region. Each region is equipped with a dedicated Low-Rank Adaptation (LoRA) module, which is trained only on the region's updated data. During inference, RAIE routes each user sequence to its corresponding region and activates the region-specific adapter for prediction. Experiments on two benchmark datasets under a time-sliced protocol that segments data into Set-up (S), Finetune (F), and Test (T) show that RAIE significantly outperforms state-of-the-art baselines while effectively mitigating forgetting. These results demonstrate that region-aware editing offers an accurate and scalable mechanism for continual adaptation in dynamic recommendation scenarios. Our code is available at https://github.com/fengaogao/RAIE.

Problem

• preference drift
• incremental update
• catastrophic forgetting
• recommendation systems
• non-stationary interactions

Innovation

• Region-Aware Editing
• Incremental Learning
• LoRA
• Preference Drift
• Catastrophic Forgetting