---
格式版本: 2
标题: "TGR: Advancing Industrial Recommendation from Generative-Paradigm Ranking toward Unified Generation and Reasoning"
原文链接: "https://arxiv.org/abs/2609.00986"
发布日期: "2026-09-01"
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\\]** Tue, 1 Sep 2026 09:41:26 UTC (2,879 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: "2026-09-02T18:53:07+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: "2026-09-02T18:51:39+08:00"
入库时间: "2026-09-02T10:53:07.225Z"
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=Scale-up&searchtype=all"
匹配关键词:
  - "Scale-up"
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: 36
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "正文主线是腾讯生成式推荐框架TGR，新增CCFormer、BARGE、HiGR和TGR-Reason及线上A/B测试、全量上线效果；来源为可追溯的arXiv论文摘要页，但未体现同行评审。固定知识库未发现同一方案，不能据此认定首次出现。文中技术与部署指标均服务于推荐、广告和内容消费，未提供可独立复用的机架级通信、网络、调度、RAS、供电或液冷机制，命中“应用与模型效率”硬否决项，不具备超节点业务准入价值。"
AI质检模型: "gpt-5.6-sol"
AI质检时间: "2026-09-02T18:54:40+08:00"
AI主题相关性: 0
AI来源权威性: 10
AI新颖性: 13
AI技术细节: 2
AI商业部署信号: 3
AI完整性: 8
AI评分提示词版本: "v17-精简生产版"
AI评分提示词SHA256: "48fb9777f386026761b4873eaff30807694fb11e9b352d7c69bf2dfde750cc7d"
AI评分知识库版本: "knowledge_base_v1-20260819+runtime.81"
AI评分知识库SHA256: "3b93d12e47749b3512f545f51c44c011bdc0931677c2cfe61e4df59a2b1a5a48"
AI评分知识库检索词: "[\"Scale-up\",\"TGR\",\"PDF\",\"arxiv.org/pdf/2609.00986\",\"HTML\",\"arxiv.org/html/2609.00986v1\",\"v1\",\"TGR-GenRank\",\"CCFormer\",\"TGR-GenRec\",\"BARGE\",\"IDs\"]"
AI评分知识库命中: "[{\"id\":\"runtime-569639751a0dbe7ef3ffccc5\",\"title\":\"[2608.17503] Predict Before Replay: Joint FEC and Flight Control for Reliable Scale-Up Links\",\"sourceType\":\"ai_excellent_article\",\"time\":\"2026-08-18\",\"matchedTerms\":[\"Scale-up\",\"PDF\",\"HTML\",\"v1\"],\"rank\":-13.013405860992304},{\"id\":\"july-correct-0111\",\"title\":\"StrataCL: Fabric-Native Communication Library for Production Supernodes\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"PDF\",\"HTML\",\"v1\"],\"rank\":-11.543771882503226},{\"id\":\"july-correct-0010\",\"title\":\"Schneider Electric and AMD release first Helios platform reference design to accelerate AI Factory deployment\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"PDF\",\"v1\",\"IDs\"],\"rank\":-8.41606444929818},{\"id\":\"july-correct-0062\",\"title\":\"Why AI Racks Need an Open Signal Conditioning Standard\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Scale-up\",\"PDF\",\"v1\"],\"rank\":-7.511302190484807},{\"id\":\"july-correct-0115\",\"title\":\"锚定 300kW 整机柜演进方向 OAII 社区三项规范联合发布，树立AIDC基础设施建设标准答案\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"HTML\",\"v1\"],\"rank\":-7.216250453347488}]"
AI摘要: "腾讯提出TGR工业推荐框架，融合生成式排序（CCFormer）、整表生成（HiGR）与语义ID推理（TGR-Reason），已在腾讯多个场景服务数亿用户；"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-02T17:13:19.543Z"
采集批次: "2026年9月2日18点28分48秒"
采集批次ID: "20260902-182848-894"
去重键: "https://arxiv.org/abs/2609.00986"
---

## Computer Science > Information Retrieval

## Title:TGR: Advancing Industrial Recommendation from Generative-Paradigm Ranking toward Unified Generation and Reasoning

[View PDF](https://arxiv.org/pdf/2609.00986) [HTML (experimental)](https://arxiv.org/html/2609.00986v1)

> Abstract:Industrial recommender systems typically rely on cascaded retrieval, pre-ranking, ranking, and reranking stages, whose separately optimized models limit scaling, fragment decision making, and lack semantic knowledge and reasoning. We present TGR (Tencent Generative Recommendation), an industrial framework that advances recommendation toward the generative paradigm along three coupled directions. TGR-GenRank upgrades ranking through CCFormer, which combines unified feature tokenization, a scalable Transformer backbone, feature-field separated cross attention, subspace token mixing, and hierarchical sequence compression while retaining per-item multi-task outputs. TGR-GenRec explores end-to-end generation under two paradigms: BARGE bridges item-boundary loss and semantic drift in hierarchical semantic-ID generation through item context-aware attention, hierarchical path reranking, and orthogonal dual-path decoding; HiGR performs whole-slate generation with prefix-structured semantic IDs, coarse-to-fine decoding, and listwise multi-objective alignment. TGR-Reason injects offline-generated semantic-ID reason tokens into online decoding, providing reasoning without request-time rollout. TGR is deployed across Tencent production surfaces serving hundreds of millions of users. CCFormer delivers significant gains in five A/B-tested scenarios and is fully launched in two, including +3.57% CTR and +1.71% advertising revenue. BARGE improves Hit@5 by 10.2-16.9% and yields +0.60% CTR and +1.70% reading time after full rollout. HiGR improves offline slate quality by 15.9-21.3% with a 5x inference speedup and achieves up to +1.22% watch time and +1.73% video views. TGR-Reason raises cold-start new-user Hit@1 by 477.8% and delivers +1.75% effective consumption and +13.09% new-user exposure-to-conversion online.

| Subjects: | Information Retrieval (cs.IR) |
| --- | --- |
| Cite as: | [arXiv:2609.00986](https://arxiv.org/abs/2609.00986) \[cs.IR\] |
|  | (or [arXiv:2609.00986v1](https://arxiv.org/abs/2609.00986v1) \[cs.IR\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2609.00986](https://doi.org/10.48550/arXiv.2609.00986) |

## Submission history

From: Yudong Li \[[view email](https://arxiv.org/show-email/31b02069/2609.00986)\]  
**\[v1\]** Tue, 1 Sep 2026 09:41:26 UTC (2,879 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2609.00986) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))