arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.03091cs.IRcs.AI
收藏

BAHSD:通过自适应蒸馏弥合黑盒序列推荐中的长尾差距

BAHSD: Bridging the Long-tail Gap via Adaptive Distillation in Black-box Sequential Recommendation

Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China(中国科学院信息工程研究所,北京,中国);School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China(中国科学院大学网络安全学院,北京,中国);Beijing Institute for General Artificial Intelligence, Beijing, China(北京一般人工智能研究院,北京,中国)

Xi Zhou, Famin Wu, Mingming Li, Hongyue Zhang, Jiao Dai, Jizhong Han, Tao Guo

AI总结:

针对黑盒序列推荐中长尾分布导致的信号异质性,提出BAHSD框架,利用多尺度一致性探测机制量化信号可靠性,并设计自适应分层目标(动态温度KL散度、排序一致性和InfoNCE对比学习)来缓解偏好固化并增强噪声鲁棒性,在尾用户上提升80%以上。

摘要:

序列推荐系统被广泛采用,但通常作为黑盒API部署,这推动了近期对模型提取的兴趣,以在本地复制其能力。然而,长尾分布导致了严重的信号异质性:密集的头部序列触发教师偏好的固化,使提取偏向局部模式,而稀疏的尾部序列产生平坦且嘈杂的预测。现有的一刀切式提取忽略了这种差异,导致噪声过拟合和次优的知识迁移。我们提出BAHSD,一种黑盒自适应蒸馏框架,通过多尺度一致性探测机制隐式量化信号可靠性来处理信号异质性。基于此,设计了自适应分层目标:动态温度KL散度缓解高置信度信号的偏好固化,而排序一致性和InfoNCE对比学习为低置信度信号提供噪声鲁棒的增强。BAHSD持续优于基线,在教师模型上获得高达4.98%的提升,在尾用户上提升80%以上,为高保真黑盒推荐提取提供了一种即插即用的解决方案。

英文摘要:

Sequential recommendation systems are widely adopted but often deployed as black-box APIs, which has driven recent interest in model extraction to replicate their capabilities locally. However, the long-tail distribution induces severe signal heterogeneity: dense head sequences trigger the solidification of teacher preference, biasing extraction toward local patterns, while sparse tail sequences yield flat, noisy predictions. Existing one-size-fits-all extraction overlooks this disparity, resulting in noise overfitting and suboptimal knowledge transfer. We propose BAHSD, a black-box adaptive distillation framework that handles signal heterogeneity via a multi-scale consistency probing mechanism to implicitly quantify signal reliability. Based on this, an adaptive hierarchical objective is designed: dynamic-temperature KL divergence mitigates preference solidification for high-confidence signals, while ranking consistency and InfoNCE contrastive learning provide noise-robust enhancement for low-confidence signals. BAHSD consistently outperforms baselines, achieving up to 4.98\% gain over the teacher and 80\%+ improvement on tail users, offering a plug-and-play solution for high-fidelity black-box recommendation extraction.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑