arXiv cs.AI· Tianqi Liu, Kaipu Sun, Kairui Fu, Wenkai Wang, Sihao Liu, Hui Zang, Ran Chen, Tao Wang, Shengyu Zhang·· 4 小时前AI 评分34
RACE:面向高效序列推荐的排序不确定性感知容量自适应定制框架
RACE: Ranking-Uncertainty Aware Capacity-Adaptive Customization for Efficient Sequential Recommendation
AI 导读
提出RACE框架,针对序列推荐中模型压缩“一刀切”导致的容量与复杂度不匹配问题,动态调整模型结构与输入序列。RACE从top-K预测的分布稳定性导出排序不确定性指标,经自监督训练的轻量预测器据其对实例进行参数与数据的个性化精简。在四个数据集上的实验显示,RACE在准确率、效率和适应性上均优于现有SOTA基线,论文已被WSDM 2027接收。
正文 · 原文
Abstract:Efficient sequential recommendation has emerged as a promising paradigm for providing real-time personalized service. To deploy models under resource-constrained environment, model compression becomes a necessity, yet existing methods typically apply uniform compression ratios, overlooking the heterogeneity in user intent complexity. This "one-size-fits-all" approach leads to a fundamental Capacity-Complexity Mismatch, where computational resources are either wasted on simple instances or insufficient for complex ones. To address this, we propose RACE (Ranking-uncertainty Aware Capacity-Adaptive Customization for Efficient Sequential Recommendation), a framework that dynamically adapts both model structures and input sequences with uncertainty-aware compression ratios. To ensure capacity-adaptive inference, RACE captures the model's intrinsic uncertainty of its predictions and transforms it into a proactive resource allocator. Specifically, we derive a list-wise ranking-uncertainty metric from the distributional stability of top-$K$ predictions, which serves as a measurable proxy for instance complexity. A lightweight predictor is then trained via self-supervision to anticipate this uncertainty from early-layer features. Guided by this signal, RACE performs instance-specific slimming in two dimensions: a hypernetwork-style channel sensitivity extractor identifies parameter redundancy, while importance-aware progressive sequence filtering mitigates data redundancy. Extensive experiments on four popular datasets demonstrate that RACE outperforms state-of-the-art baselines on accuracy, efficiency, and adaptability.
| Comments: | Accepted by WSDM 2027 (Main Conference) |
| Subjects: | Information Retrieval (cs.IR); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.18244 [cs.IR] |
| (or arXiv:2611.18244v1 [cs.IR] for this version) | |
| https://doi.org/10.48550/arXiv.2611.18244 arXiv-issued DOI via DataCite (pending registration) |
|
| Related DOI: | https://doi.org/10.1145/3862322.3863404
DOI(s) linking to related resources |
Submission history
From: Tianqi Liu [view email]
[v1]
Fri, 13 Nov 2026 14:40:26 UTC (9,930 KB)
来源:arXiv cs.AI · arxiv.org