Search papers, labs, and topics across Lattice.
This paper introduces SharpRec, a novel framework for LLM-based Cross-Domain Sequential Recommendation (CDSR) that addresses critical challenges such as cross-domain knowledge conflict and performance saturation during model merging. By employing Sharpness-aware Geometric Alignment and Preference Salience Activation, SharpRec stabilizes the merging process and recovers essential features, leading to enhanced performance in multi-domain scenarios. Extensive experiments show that SharpRec significantly outperforms existing state-of-the-art methods, confirming its effectiveness in integrating diverse knowledge sources for improved recommendation accuracy.
Merging LLMs can lead to performance saturation, but SharpRec's innovative approach recovers essential features to consistently outperform state-of-the-art methods.
LLM-based Cross-Domain Sequential Recommendation (CDSR) leverages LLMs to enhance target performance via deep semantic reasoning, alleviating the dependency on overlapping users. Among LLM-based paradigms, model merging is particularly promising for multi-domain scenarios due to its superior scalability and flexibility in integrating diverse knowledge sources. However, our empirical investigations reveal two critical bottlenecks: (1) cross-domain knowledge conflict; and (2) performance saturation in multi-domain fusion. Our analysis attributes these phenomena to parameter-level misalignment and statistical homogenization during the merging process. To address these bottlenecks, we propose SharpRec, Sharpness-aware Model Merging with Salience Recovery for LLM-based CDSR, a framework designed to lift the performance upper bound of merged models. SharpRec incorporates two synergistic modules: Sharpness-aware Geometric Alignment to establish a stable geometric foundation for interference-free fusion; and Preference Salience Activation to effectively recover the distinctive features essential for bolstering target domain performance. Extensive experiments in both dual-domain and multi-domain scenarios demonstrate that SharpRec consistently outperforms state-of-the-art baselines.
