快手提出生成式强化学习出价技术，驱动在线广告收入显著提升

全景网2025-09-25

近日，快手首次提出生成式强化学习出价技术，并将其应用于大规模广告系统中。该项技术显著提升了广告竞价的智能化水平，实现了出价策略从传统“依赖单步状态决策”到新一代“序列决策”的关键跨越，在为广告主带来更高投放回报的同时，也提升了平台整体收益，真正实现了双赢。2025年至今，该技术已成功推动快手广告收入提升3%，成效显著。广告出价：实时竞价中的“智能大脑”在实时竞价(RTB)广告系统中，出价模块作为...

网页链接

免责声明：本文观点仅代表作者个人观点，不构成本平台的投资建议，本平台不对文章信息准确性、完整性和及时性做出任何保证，亦不对因使用或信赖文章信息引发的任何损失承担责任。

精彩评论

我们需要你的真知灼见来填补这片空白

发表看法

{"i18n":{"language":"zh_CN"},"isChannel":false,"data":{"share":"https://www.laohu8.com/m/news/2570096844?lang=zh_CN&edition=full","thumbnail":"","is_english":false,"pubTime":"2025-09-25 18:34","share_image_url":"https://static.laohu8.com/9a95c1376e76363c1401fee7d3717173","id":"2570096844","market":"hk","top_or_hot":-1,"title":"快手提出生成式强化学习出价技术，驱动在线广告收入显著提升","media":"全景网","content":"<div>\n<p>近日，快手首次提出生成式强化学习出价技术，并将其应用于大规模广告系统中。该项技术显著提升了广告竞价的智能化水平，实现了出价策略从传统“依赖单步状态决策”到新一代“序列决策”的关键跨越，在为广告主带来更高投放回报的同时，也提升了平台整体收益，真正实现了双赢。2025年至今，该技术已成功推动快手广告收入提升3%，成效显著。广告出价：实时竞价中的“智能大脑”在实时竞价(RTB)广告系统中，出价模块作为...</p>\n\n<a href=\"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202509251835059544800a&s=b\">网页链接</a>\n\n</div>\n","source":"tencent","html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>快手提出生成式强化学习出价技术，驱动在线广告收入显著提升</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 11px; color: #7E829C; margin: 0;line-height: 11px;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\n快手提出生成式强化学习出价技术，驱动在线广告收入显著提升\n</h2>\n\n<h4 class=\"meta\">\n\n\n2025-09-25 18:34 北京时间&nbsp;&nbsp;&nbsp;<a href=http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202509251835059544800a&s=b><strong>全景网</strong></a>\n\n\n</h4>\n\n</header>\n<article>\n<div>\n<p>近日，快手首次提出生成式强化学习出价技术，并将其应用于大规模广告系统中。该项技术显著提升了广告竞价的智能化水平，实现了出价策略从传统“依赖单步状态决策”到新一代“序列决策”的关键跨越，在为广告主带来更高投放回报的同时，也提升了平台整体收益，真正实现了双赢。2025年至今，该技术已成功推动快手广告收入提升3%，成效显著。广告出价：实时竞价中的“智能大脑”在实时竞价(RTB)广告系统中，出价模块作为...</p>\n\n<a href=\"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202509251835059544800a&s=b\">网页链接</a>\n\n</div>\n\n\n</article>\n</div>\n</body>\n</html>\n","isBrief":false,"type":0,"news_type":1,"symbol":"LU0463099449.HKD","symbol_name":"SCHRODER ISF CHINA OPPORTUNITIES \"A\" (HKD) ACC","start_time":0,"source_url":"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202509251835059544800a&s=b","article_id":"2570096844","we_media_id":null,"thumbnails":[],"rights":{"source":"tencent","url":"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202509251835059544800a&s=b","rn_cache_url":null,"customStyle":"body{padding-top:10px;}#news_title{font-weight:bold;#titleStyle#;}#news_description span{font-size:12px;#descriptionStyle#;}.footer-note{#statement#}","selectors":".mod-LoadTzbdNews, body","filters":".relate-stock, .hot-list, .recom-box, .wx-sou","directOrigin":true},"url":"https://stock-news.laohu8.com/highlight/detail?id=2570096844","pubTimestamp":1758796440,"columns":[],"sourceInfo":{"source_id":"tencent","name":"腾讯"},"weMediaInfo":null,"summary":"近日，快手首次提出生成式强化学习出价技术，并将其应用于大规模广告系统中。面对这一挑战，快手创新融合两类技术，提出生成式强化学习出价技术——在继承生成模型序列建模能力的基础上，引入强化学习的目标导向优化机制，实现出价策略在长序列决策中的新突破。2025年至今，生成式强化学习出价技术已在快手广告系统中全面部署，并推动平台广告收入提升3%。","collect":0,"end_time":0,"defaultTopTitle":"qq.com","property":[],"viewcount":null,"language":"zh","relate_stocks":{"81024":"快手-WR","LU0463099449.HKD":"SCHRODER ISF CHINA OPPORTUNITIES \"A\" (HKD) ACC","LU2097828805.USD":"AZ EQUITY CHINA \"A-AZ\" (USD) ACC","LU0359201885.HKD":"BGF CHINA \"A2\" (HKD) ACC","LU0348766576.USD":"ALLIANZ LITTLE DRAGONS \"A\" (USD) INC","LU0211977185.USD":"EASTSPRING INVESTMENTS GREATER CHINA EQUITY \"A\" ACC","LU0348767384.USD":"ALLIANZLITTLE DRAGONS \"A\" (USD) ACC","LU0140636845.USD":"施罗德大中华区股票A Acc","LU0307460666.USD":"EASTSPRING INVESTMENTS CHINA EQUITY \"A\" ACC","LU2097828474.EUR":"AZ EQUITY CHINA \"A\" (EUR) ACC A","LU0588546209.SGD":"Eastspring Investments - China Equity Fund AS SGD","LU0593848301.USD":"未来资产亚洲卓越消费股票基金A","LU1188198961.HKD":"SCHRODER ISF CHINA OPPORTUNITIES \"A\" (HKD) INC QV","01024":"快手-W","LU0244354667.USD":"SCHRODER ISF CHINA OPPORTUNITIES \"A\" ACC","LU0359201612.USD":"贝莱德中国基金A2","LU1251922891.USD":"NINETY ONE GSF ALL CHINA EQUITY \"A\" (USD) ACC","LU0359202008.SGD":"Blackrock China Fund A2 SGD-H","LU0051755006.USD":"摩根大通中国A (dist)","BK1137":"广告","LU1023057109.AUD":"BGF CHINA \"A2\" (AUDHDG) ACC","BK1610":"ETF&股票定投概念","LU1720050803.USD":"安联全方位中国股票基金","LU0326950275.SGD":"Schroder ISF China Opportunities A Acc SGD-H","LU0210526637.USD":"JPM CHINA \"A\" (USD) ACC","LU0456827905.SGD":"JPMorgan Funds - China A (acc) SGD","BK1615":"港股-互联网","BK1095":"互动媒体与服务","BK1575":"同股不同权","LU1303224171.USD":"NINETY ONE GSF ALL CHINA EQUITY \"A\" (USD) INC","LU1719994722.HKD":"NINETY ONE GSF ALL CHINA EQUITY \"A\" (HKD) ACC","LU2097828557.USD":"AZ EQUITY CHINA \"A\" (USD) ACC","LU2097828631.EUR":"AZ EQUITY CHINA \"A\" (EUR) ACC","BK1608":"元宇宙概念","LU1770034418.SGD":"ALL CHINA EQUITY \"A\" (SGDHDG) ACC","LU1794554557.SGD":"Allianz All China Equity AT Acc H2-SGD","BK1591":"就地过年概念","LU2097828714.EUR":"AZ EQUITY CHINA \"BAZ\" (EUR) ACC","01163":"虎视传媒","BK1590":"短视频概念股"},"translate_title":"Kuaishou proposes generative reinforcement learning bidding technology to drive a significant increase in online advertising revenue","themeId":null,"isJumpTheme":false,"ttsUrl":null,"symbols_score_info":{"81024":0.6,"01024":1.5,"01163":1.5,"KSTmain":0.6},"content_text":"近日，快手首次提出生成式强化学习出价技术，并将其应用于大规模广告系统中。该项技术显著提升了广告竞价的智能化水平，实现了出价策略从传统“依赖单步状态决策”到新一代“序列决策”的关键跨越，在为广告主带来更高投放回报的同时，也提升了平台整体收益，真正实现了双赢。2025年至今，该技术已成功推动快手广告收入提升3%，成效显著。广告出价：实时竞价中的“智能大脑”在实时竞价(RTB)广告系统中，出价模块作为连接广告主与流量的核心枢纽，需要在多重约束条件下实现广告投放效果的最优化。这一过程面临三大挑战。首先是既要控制支出，又要追求效益。广告主不仅需要将单日花费严格控制在预算范围内，还应力求降低每次转化的成本，实现成本效益的双重优化。其次是，未来环境的不确定性。系统难以预知即将到来的流量状况与竞争对手行为，因此必须依赖实时花费与成本等数据，动态调整出价策略以应对波动市场。最后是决策之间的连锁影响。每一次出价不仅直接影响广告的展示与消耗，还会改变账户状态(如剩余预算)，进而干扰后续出价决策，形成一个连续且复杂的序列决策问题。三代技术革新，快手提出新一代生成式强化学习出价技术快手出价技术已历经三代演进：从PID控制——基于当前状态与目标的偏差进行调节，响应直接但适应性有限，难以应对复杂多变的竞价环境；到第二代MPC预测模型——能够预测短期未来并优化动作，但预测范围有限，易陷局部最优；再到第三代强化学习——通过历史数据学习不同状态下的最优出价策略，最大化长期收益，兼顾效果与安全性。近年来，尽管强化学习在自动出价中表现优异，但其决策仍依赖单步状态信息，对序列历史利用不足。而生成式模型擅长对序列数据进行建模和生成，却不擅长直接优化目标收益。面对这一挑战，快手创新融合两类技术，提出生成式强化学习出价技术——在继承生成模型序列建模能力的基础上，引入强化学习的目标导向优化机制，实现出价策略在长序列决策中的新突破。两大核心算法实现广告出价“多维思考”，助推智能决策升级在将生成模型应用于强化出价策略的过程中，主要面临两大挑战：一是对高质量数据集的依赖较强，限制了泛化能力；二是生成模型难以直接优化序列整体收益，导致出价策略与广告主全局目标存在偏差。针对这些挑战，快手提出了GAVE和CBD两大创新算法。GAVE算法融合了Score-based RTG机制与价值函数引导的动作探索策略，显著提升了离线训练中的探索能力，该成果助力团队荣获NeurIPS 2024自动出价竞赛双赛道冠军。CBD算法则聚焦于生成模型与优化目标的对齐问题，引入“Completer(扩散补全)”和“Aligner(偏好对齐)”双模块，通过补全未来状态轨迹并进行校准，使出价策略更贴合广告主真实需求，提升了系统长程规划与决策可解释性。2025年至今，生成式强化学习出价技术已在快手广告系统中全面部署，并推动平台广告收入提升3%。在多项关键客户指标上实现突破——如在成本控制场景中，消耗提升2.0%，CPA达标率提升1.9%。展望未来，快手将围绕两大方向持续演进。一是构建出价基座大模型，整合多场景数据训练通用模型，发挥规模效应；二是发展出价推理大模型，引入大语言模型增强推理能力，提升可解释性与复杂决策水平。快手将继续推动自动出价技术向更高层次的智能化迈进，为数字营销行业注入创新动力。","kind":"news","is_publish_news":true,"is_publish_highlight":false,"is_publish_live":false,"is_publish_wemedia":null,"editions":null,"column":"","sentiment":"1","news_tag":"","news_rank":0,"symbols":[],"gpt_button":0,"need_auth":false,"code":"91000000","status":"200"},"commentList":[],"isCommentEnd":true,"newsSizeData":{"likeSize":0,"commentSize":0,"repostSize":0,"favoriteSize":0,"likeStatus":false,"favoriteStatus":false},"APP":{"userAgent":"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)","isDev":false,"isTTM":false,"tenantId":"TBCN","deviceId":"web-server-community-laohu8-v3","version":"4.36.2","shortVersion":"4.36.2","platform":"web","vendor":"web","appName":"laohu8","isIOS":false,"isAndroid":false,"isTiger":false,"isTHS":false,"isWeiXin":false,"isWeiXinMini":false,"isWeiBo":false,"isQQ":false,"isBaiduSwan":false,"isBaiduBox":false,"isDingTalk":false,"isToutiao":false,"isOnePlus":false,"isHuaWei":false,"isXiaomi":false,"isXiaomiWebView":false,"isOppo":false,"isVivo":false,"isSamsung":false,"isMobile":false},"href":"/m/news/2570096844","isCrawlerRequest":true}