两篇新研究论文解决了推荐系统中超长用户行为序列建模的挑战。第一篇论文SequenceO1介绍了一个在抖音部署的端到端框架，该框架使用Sketch Attention和Stacked Target-to-History Cross Attention来压缩和推理长达10万次交互的历史记录。第二篇论文提出了一个两阶段框架，通过修改评分机制和微调特定组件，在不依赖缓存历史记录的情况下缩小长短视图的性能差距。 AI
影响 这些方法旨在通过使推荐系统能够处理和学习更长的用户交互历史来提高其效率和有效性。
排序理由 两篇在arXiv上发表的学术论文，提出了推荐系统序列建模的新方法。
在 arXiv cs.IR (Information Retrieval) 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2609.08443v1 Announce Type: cross Abstract: Modeling long-term user behavior is central to sequential recommendation and billion-scale industrial recommender systems, yet production ranking models operate under strict latency, memory, communication, and training-throughput …
Modeling long-term user behavior is central to sequential recommendation and billion-scale industrial recommender systems, yet production ranking models operate under strict latency, memory, communication, and training-throughput constraints. At the 100K scale, the challenge exte…
Sequential recommenders are typically trained on long user histories to capture rich behavioral signals, yet serving with training-length sequences is often impractical due to real-time efficiency constraints. Directly using only recent behaviors leads to a severe performance dro…