通过组件级蒸馏提升序列推荐泛化能力，解决稀疏历史下的模型过拟合问题。
GOD: Enhancing Generalization via Deep Grafting for Sequential Recommendation
- 将教师模型的组件替换为可训练学生组件，构建混合模型进行反馈
- 在三个真实数据集上性能领先基线最高13.92%
- 推理时仅用学生模型，无额外计算开销，适合工业部署
序列推荐器常因历史记录稀疏和噪声而泛化能力受限。知识蒸馏通过将教师模型的密集监督传递给学生模型缓解此问题。然而，现有方法多独立运行师生模型，仅匹配输出或表征，导致学生各组件的影响相互纠缠，难以区分泛化弱是源于不可靠嵌入、过拟合编码，还是对稀疏历史的共适应。本文提出嫁接导向蒸馏（GOD），一种基于组件级蒸馏的框架，通过替换教师模型中选定的冻结组件为可训练的学生组件，构建混合源模型。GOD利用这些混合模型评估学生嵌入（使用教师编码器）与学生编码器（使用教师嵌入），实现组件级反馈。推理阶段仅使用学生模型，无额外开销。在三个真实世界数据集上，GOD性能优于当前最优基线，最高提升达13.92%。
原文摘要 · Abstract (English)
Sequential recommenders often struggle with sparse and noisy histories, limiting generalization to unseen interactions. Knowledge distillation mitigates this by transferring dense supervision from a teacher to a student. However, most distillation methods run teacher and student independently, then match student outputs or representations to the teacher. Such supervision entangles student-component effects, blurring whether weak generalization stems from unreliable embeddings, overfitted encoding, or co-adaptation to sparse histories. In this paper, we propose Graft-Oriented Distillation (GOD), a component-level distillation framework for improved generalization through grafting. Grafting denotes replacing selected frozen-teacher components with trainable student counterparts to build hybrid source models. GOD uses these hybrid models to evaluate student embeddings with the teacher encoder and the student encoder with teacher embeddings, providing component-level feedback. At inference, GOD uses only the student, incurring no additional cost. Across three real-world datasets, GOD outperforms state-of-the-art baselines by up to 13.92%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。