刘琼 腾讯研究院产业研究中心主任 徐一平 腾讯研究院高级研究员 人工智能生成 广告的技术演进与模式变革 人工智能，尤其是以大语言模型和多模态模型为代表的生成式智能，正在全球范围内重塑广告系统的底 层逻辑。当前，数字广告技术已经从上一代的程序化广告推荐系统，向能够理解用户意图、分析用户情 感和行为的智能广告系统演进。广告不再仅是页面中的插图或文字，逐步演化为由人工智能"理解用户 需求、生成内容、决定投放"的完整闭环。不仅如此，生成式智能技术在广告创意生成、智能投放、智 能审核等环节都有广泛应用。以谷歌为例，2025年I/O大会[ I/O大会 （Innovatio n in t he Open） 即 Google开发者大会，于2025年8月13日召开，聚焦 Google 创新技术，Android、AI、web、cloud 等领域 最新动态。]上发布的"AI模式"将Gemini模型深度嵌入搜索、推荐与广告生成流程中，展示了广告从"被 动呈现"到"主动交互"的彻底转变。腾讯AI广告创意平台——"妙思"借自主研发的混元大模型打通创意 制作、投放流程与广告审核多个环节。 广告的投放逻辑发生质变。如在跨境电商广告投放中，生 ...

作者 | AICon 全球人工智能开发与应用大会 策划 | 罗燕珊 编辑 | 宇琪 当大模型从通用技术探索深入产业场景，搜索、广告与推荐系统作为连接用户需求与业务价值的 核心链路，正迎来全链路智能重构。那么，生成式推荐真正落地后的关键挑战是什么？又应该如 何解决？ 近日 InfoQ《极客有约》X AICon 直播栏目特别邀请了 京东内容推荐架构负责人颜林 担任主持 人，和 荣耀 AI 算法专家冯晓东、京东算法总监张泽华、中科大计算机学院副教授王皓 一 起，在 AICon 全球人工智能开发与应用大会 2025 北京站 即将召开之际，共同探讨生成式推 荐的落地洞察。 部分精彩观点如下： 完整直播回放可查看： https://www.infoq.cn/video/0ViWrdqyQwNvO7TdQpyD 以下内容基于直播速记整理，经 InfoQ 删减。 行业真正做到端到端的统一 pipeline 仍有较大差距，更多工作还是在 pipeline 的单点与大模型 结合。 搜广推场景中的 scaling law 依然成立，并且仍在快速上升阶段。 低价值场景用小模型覆盖，高价值场景用大模型争取额外收益。 不应拘泥于某项技术 ...

Group 1 - The core theme of the live broadcast is the practical challenges and advancements of search, recommendation, and advertising systems in the era of large models [2][4][7] - Experts from companies like Honor, Huawei, and JD.com will discuss the evolution and difficulties faced by search and advertising systems with the integration of large models [2][4][7] - Key challenges include scaling generative recommendations, the effectiveness of scaling laws in search and advertising, balancing online inference latency and costs, and integrating multimodal and behavioral large models throughout the entire process [2][4][7] Group 2 - The live broadcast is scheduled for November 26, from 20:00 to 21:30, hosted by Yan Lin, the content recommendation architecture leader at JD.com [3] - The event will feature experts such as Feng Xiaodong from Honor, Wang Hao from the University of Science and Technology of China, and Zhang Zehua from JD.com, focusing on the full-chain upgrade in recommendation and advertising [3][4] - The broadcast will cover practical insights into technical architecture, application cases, and engineering deployment related to large models, providing valuable information for various industries [5][7]

近年来，在推荐系统领域，传统 "召回 + 排序" 级联式架构的收益正逐渐触顶，而 ChatGPT 等 大语言模型 则展现了强大的涌现能力和符合 Scaling Law 的巨大潜力 —— 这股变革性的力量使 "生 成式推 荐" 成为当下最热门的话题之一。不同于判别式模型孤立地计算用户喜欢某件物品的概率，"生成式推荐" 能够利用层次化语 义 ID 表示用户历史行为序列，并基于生成式模型结构 直接生成 用户下一批可能交互的物品列表。这种推荐模式显著提升了模型的 智能上限 ，并为推荐场景引入 Scaling Law 的可能性。 快手 OneRec 的成功落地，更是彻底引爆了推荐圈子。凭借端到端的推荐大模型，重构现今的推荐系统不再是空谈，它已证明是一场 资 源可控、能带来真实线上 收益的 推荐革 命。 然而，对于这一可能革新整个推荐系统的新范式，各大厂却讳莫如深，核心技术细节与公开表现鲜有披露。开源社区与一线大厂的探索似乎正在脱钩，技术鸿沟 日渐明显。 如何破局？ 近日，中国科学技术大学 LDS 实验室何向南、王翔团队联合 Alpha Lab 张岸团队正式发布 MiniOneRec 。这一框架作为生成式推荐领域首个完整 ...

Group 1: Key Recommendations - The report highlights the strong growth potential of the company "October Rice Field" (09676) in the health food sector, driven by the launch of new products and expansion into high-potential sales channels [6] - The company has exceeded expectations in both new product sales and channel expansion, with corn products emerging as a significant growth driver alongside rice products [6] - Revenue projections for "October Rice Field" are estimated at 6.951 billion, 8.371 billion, and 9.884 billion yuan for 2025-2027, reflecting year-on-year growth rates of 21%, 20%, and 18% respectively [6] Group 2: Industry Insights - The mechanical equipment sector is experiencing a cyclical reversal and growth, emphasizing the importance of self-sufficiency and domestic substitution due to trade tensions [8] - The report suggests that the military industry is poised for growth, with increased domestic demand and potential for foreign trade expansion leading to a revaluation of the sector [10] - The telecommunications industry is expected to maintain steady growth, with significant opportunities arising from advancements in computing power and satellite internet technologies [10][11] Group 3: Investment Opportunities - The report identifies generative recommendation technology as a key area for investment, with major internet companies exploring its potential to enhance business outcomes [11] - Companies that successfully implement generative recommendation technology are expected to achieve substantial commercial value, surpassing industry averages [11] - Catalysts for growth in this area include improved advertising conversion rates and successful implementation by leading internet firms [11]

Core Insights - Alibaba Mama showcased its latest advancements in AIGR (Generative Recommendation) at the first International General Artificial Intelligence Conference, TongAI [1] Group 1 - The URM General Recall Model from the LMA2 advertising model series was prominently featured [1]