---
格式版本: 2
标题: "Decoupled Temporal Encoding for Generative Recommendation"
原文链接: "https://arxiv.org/abs/2608.16274"
发布日期: "2026-08-17"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Mon, 17 Aug 2026 08:47:21 UTC (693 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:19:46+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:12:40+08:00"
入库时间: "2026-08-20T07:19:46.539Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=delivery&searchtype=all"
匹配关键词:
  - "delivery"
  - "deployment"
  - "AI"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 8
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "论文主题为推荐系统中的时间编码，属于信息检索领域，与超节点/AI Rack/机柜级AI基础设施完全无关。命中词delivery仅指外卖配送，非数据中心交付。来源虽为arXiv，但主题不相关，判为无关。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T15:21:18+08:00"
AI主题相关性: 0
AI来源权威性: 6
AI新颖性: 2
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
AI摘要: "该论文提出解耦时间编码（DTE）框架，用于生成式推荐系统，将用户行为中的宏观时间动态与局部顺序信息分离，通过个性化宏时模块和时间门控微序模块处理就餐高峰、周末变化等规律。该方法面向外卖和即时零售场景，参数高效且易部署集成。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:24:53.793Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.16274"
---

## Computer Science > Information Retrieval

## Title:Decoupled Temporal Encoding for Generative Recommendation

Authors:[Pengfei Jia](https://arxiv.org/search/cs?searchtype=author&query=Jia,+P), [Jingjian Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+J), [Jingmao Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+J), [Ge Zhang](https://arxiv.org/search/cs?searchtype=author&query=Zhang,+G), [Feng Shi](https://arxiv.org/search/cs?searchtype=author&query=Shi,+F)

[View PDF](https://arxiv.org/pdf/2608.16274)

> Abstract:Positional encoding is a fundamental component of Transformer-based generative recommendation models, where user histories are modeled as autoregressive item sequences. Most positional encoding methods are inherited from natural language processing and mainly represent discrete item order. However, recommendation sequences go beyond ordered lists, as timestamps and temporal effects also shape item relations. Our work is motivated by a real-world food delivery and instant retail recommendation system, where user behavior exhibits multi-level temporal regularities, including recency effects, meal-time peaks, weekday-weekend shifts, and promotion-driven traffic bursts. Existing methods partially address this issue through timestamp features, interval embeddings, decay functions, or attention biases, but they usually inject heterogeneous temporal signals through a unified representation or a single modeling pathway, making it difficult to distinguish broad temporal dynamics from local order cues. To address this limitation, we propose Decoupled Temporal Encoding, a lightweight framework for generative recommendation. DTE separates temporal dynamics from order information through two complementary modules: a personalized macro-temporal module that injects compact temporal primitives into item embeddings, and a time-gated micro-sequential module that introduces relative-order bias only when interactions are temporally dense. DTE is also parameter-efficient and deployment-friendly, allowing easy integration into existing systems.

| Comments: |  |
| --- | --- |
| Subjects: | Information Retrieval (cs.IR); Artificial Intelligence (cs.AI) |
| Cite as: | [arXiv:2608.16274](https://arxiv.org/abs/2608.16274) \[cs.IR\] |
|  | (or [arXiv:2608.16274v1](https://arxiv.org/abs/2608.16274v1) \[cs.IR\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.16274](https://doi.org/10.48550/arXiv.2608.16274) |

## Submission history

From: Pengfei Jia \[[view email](https://arxiv.org/show-email/5f883b91/2608.16274)\]  
**\[v1\]** Mon, 17 Aug 2026 08:47:21 UTC (693 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.16274) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))