[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84686-en":3,"doc-seo-84686-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},84686,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","HGenPush: A Heterogeneous Generative Recommendation Architecture for Industrial Push Notification Systems","HGenPush presents an end-to-end heterogeneous generative recommendation architecture for industrial push notification systems. It targets short-video platform needs where users demand both high-quality content and trusted authors. The approach combines a hybrid multi-scenario, multi-perspective user behavior understanding module, a unified dual-branch heterogeneous generator for video and author recommendation, and a lightweight multi-token prediction strategy that avoids autoregressive semantic-ID generation. A preference alignment module uses user feedback as reward to improve output quality. Deployed at Kuaishou, it achieves a 0.181% increase in daily active users.","HGenPush: A Heterogeneous Generative Recommendation Architecture for Industrial Push Notification Systems  \nXiao Liang∗ Kuaishou Technology Beijing, China [liangxiao@kuaishou.com](liangxiao@kuaishou.com)  \nJiali Feng∗ Kuaishou Technology Beijing, China [fengjiali05@kuaishou.com](fengjiali05@kuaishou.com)  \nXin Feng  \nKuaishou Technology Beijing, China [fengxin05@kuaishou.com](fengxin05@kuaishou.com)  \narXiv :2607 .03362v 1 [ cs .IR] 3 Jul 2026  \nYiqing Wang Kuaishou Technology  \nBeijing, China [wangyiqing05@kuaishou.com](wangyiqing05@kuaishou.com)  \nZhihui Deng  \nKuaishou Technology Beijing, China [dengzhihui@kuaishou.com](dengzhihui@kuaishou.com)  \nXuanping Li† Kuaishou Technology Beijing, China [lixuanping@kuaishou.com](lixuanping@kuaishou.com)  \nBaolin Ye  \nKuaishou Technology Beijing, China [yebaolin@kuaishou.com](yebaolin@kuaishou.com)  \nCunyi Zhang  \nKuaishou Technology Beijing, China [zhangcunyi@kuaishou.com](zhangcunyi@kuaishou.com)  \nKaiqiao Zhan  \nKuaishou Technology Beijing, China [zhankaiqiao@kuaishou.com](zhankaiqiao@kuaishou.com)  \nSiyao Feng  \nKuaishou Technology Beijing, China [fengsiyao@kuaishou.com](fengsiyao@kuaishou.com)  \nHuajin Sun  \nKuaishou Technology Beijing, China [sunhuajin@kuaishou.com](sunhuajin@kuaishou.com)  \nYanan Niu  \nKuaishou Technology Beijing, China [niuyanan@kuaishou.com](niuyanan@kuaishou.com)  \nKun Gai  \nKuaishou Technology Beijing, China[gai.kun@qq.com](gai.kun@qq.com)  \nAbstract  \nWith the explosive growth of content platforms, recommendation systems need to better satisfy user demands to enhance user satisfaction and retention. Taking short-video platforms as an example, users not only seek high-quality content but also trusted authors. Although generative recommendation systems have achieved breakthroughs in recent years, existing methods primarily generate single-type recommendation content and typically employ the inefficient autoregressive paradigm to generate semantic IDs. In this paper, we propose an end-to-end heterogeneous generative recommendation architecture called HGenPush. First, we design a hybrid user behavior understanding module that integrates multiscenario and multi-perspective behaviors to capture precise user interest. Then, we design a dual-branch heterogeneous generative recommendation module that integrates video recommendation and author recommendation within a unified framework. In addition, to improve generation efficiency, we design a lightweight  \n∗ Both authors contributed equally to this research.†Corresponding author.  \nThis work is licensed under a Creative Commons Attribution 4 .0 International License. KDD’26, Jeju Island, Republic of Korea  \n© 2026 Copyright held by the owner/author(s) .  \nACM ISBN 979-8-4007-2259-2/2026/08  \n[https://doi.org/10.1145/3770855.3818429](https://doi.org/10.1145/3770855.3818429)  \nmulti-token prediction method that discards the autoregressive paradigm. Finally, we design a user consumption preference alignment module, which leverages user feedback as reward signals to guide the model toward generating higher-quality content, thereby enhancing user experience and engagement. Through these designs, HGenPush simultaneously fulfills users’ demands for high-quality content and trusted authors. We have deployed HGenPush on the push notification system of Kuaishou, a large-scale short-video platform, achieving a significant 0.181% increase in daily active users.  \nCCS Concepts  \n• Information systems → Mobile information processing systems.  \nKeywords  \nPush Notification System, Generative Recommendation, Multi-token Prediction, Preference Alignment  \nACM Reference Format:  \nXiao Liang, Jiali Feng, Xin Feng, Yiqing Wang, Baolin Ye, Siyao Feng, Zhihui Deng, Cunyi Zhang, Huajin Sun, Xuanping Li, Kaiqiao Zhan, Yanan Niu, and Kun Gai. 2026. HGenPush: A Heterogeneous Generative Recommendation Architecture for Industrial Push Notification Systems. In Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining","cbCaigrFO39P87pr","https://ap.wps.com/l/cbCaigrFO39P87pr","pdf",1236774,2,1,9,"English","en",105,"# Introduction\n## Motivation: heterogeneous interests in content platforms\n## Generative recommendation background\n## Challenges and proposed solution","[{\"question\":\"What problem does HGenPush address in industrial push notification systems?\",\"answer\":\"HGenPush addresses the need to recommend both high-quality content and trusted authors, especially where user interests are heterogeneous across item types like videos and authors.\"},{\"question\":\"How does HGenPush model user behavior and preferences?\",\"answer\":\"It uses a hybrid user behavior understanding module that integrates multi-scenario and multi-perspective behaviors to capture precise user interests, then aligns generation with user feedback as reward signals.\"},{\"question\":\"How does HGenPush improve generation efficiency compared with autoregressive methods?\",\"answer\":\"It introduces a lightweight multi-token prediction method that discards the autoregressive paradigm used to generate semantic IDs, improving generation efficiency.\"}]",1784197655,23,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"hgenpush-a-heterogeneous-generative-recommendation-architecture-for-industrial-push-notification-systems","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,47,50],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":20},"https://docshare.wps.com/document/","Document",{"item":48,"name":12,"@type":43,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":43,"position":52},"https://docshare.wps.com/document/hgenpush-a-heterogeneous-generative-recommendation-architecture-for-industrial-push-notification-systems/84686/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-23","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does HGenPush address in industrial push notification systems?","Question",{"text":75,"@type":76},"HGenPush addresses the need to recommend both high-quality content and trusted authors, especially where user interests are heterogeneous across item types like videos and authors.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does HGenPush model user behavior and preferences?",{"text":80,"@type":76},"It uses a hybrid user behavior understanding module that integrates multi-scenario and multi-perspective behaviors to capture precise user interests, then aligns generation with user feedback as reward signals.",{"name":82,"@type":73,"acceptedAnswer":83},"How does HGenPush improve generation efficiency compared with autoregressive methods?",{"text":84,"@type":76},"It introduces a lightweight multi-token prediction method that discards the autoregressive paradigm used to generate semantic IDs, improving generation efficiency.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]