Search papers, labs, and topics across Lattice.
This paper introduces SelfDR, a Self-Distillation from Reasoning framework designed to enhance the efficiency of LLM-based recommendation systems by distilling reasoning-enhanced predictions directly from the same model. By training a reasoner to generate targeted rationales that inform a student recommender, the approach effectively balances improved recommendation accuracy with reduced computational costs. Extensive experiments across three public datasets demonstrate that SelfDR not only boosts recommendation effectiveness but also maintains inference efficiency, making it suitable for real-world applications.
SelfDR achieves high-quality recommendations by distilling reasoning from the same LLM, cutting computational costs without sacrificing performance.
Large Language Models (LLMs) have recently emerged as powerful backbones for recommendation. To better elicit their capabilities, reasoning has been widely incorporated to help LLMs interpret rich textual signals and improve recommendation accuracy. However, explicitly generating intermediate reasoning traces often incurs substantial computational costs, which limits practical deployment in real-world recommender systems. To address this challenge, we propose SelfDR, a Self-Distillation from Reasoning framework for LLM-based Recommendation. SelfDR distills an LLM's own reasoning-enhanced predictions to produce recommendations directly, improving recommendation effectiveness while maintaining inference efficiency. All components in the framework are built on the same base LLM, without relying on any external models. Specifically, the teacher recommender is constructed by training a reasoner with downstream performance as the reward, enabling it to generate targeted rationales that are later incorporated into the teacher's input. A student recommender for direct recommendation, with the same underlying model, then learns from the teacher through self-distillation with a dynamic weighting strategy. Extensive experiments on three public datasets validate the effectiveness, rationality, and efficiency of SelfDR. Codes are available at https://github.com/JiangDeccc/SelfDistillation.
