f9yx0du 发表于 2024-9-4 01:21:29

他的UTD科研证明,AI算法能够给数字营销带来“巨变”……| 学术对话No.1


    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">倘若</span>一个商场<span style="color: black;">连续</span>促销,你会有立即购物的冲动吗?可能不会!<span style="color: black;">无</span>间隔的促销或许会降低你的消费激情。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">倘若</span>一个线上购物平台<span style="color: black;">连续</span>的向你推出同一品类的东西,你会买账吗?可能不会!过于<span style="color: black;">类似</span>的商品会让你产生审美疲劳,并失去下单的欲望。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">然而,你相信吗,<span style="color: black;">针对</span>这些问题的改进,不需要商家或品牌方,仅仅<span style="color: black;">经过</span>人工智能就<span style="color: black;">能够</span>实现!从语音识别到图像处理、从自动驾驶汽车到医疗<span style="color: black;">设备</span>人,如今再到营销<span style="color: black;">行业</span>,人工智能的<span style="color: black;">“深度强化学习”</span>为数字<span style="color: black;">行业</span>带来<span style="color: black;">非常多</span>前所未有的机会。</p><img src="https://mmbiz.qpic.cn/mmbiz_jpg/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOFiaKucZj2d3VT1k8Vaicbj62bAF5jsWhJKneSfeficmwjp07ib7VDmZKOg/640?wx_fmt=jpeg&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">照片</span><span style="color: black;">源自</span>:©千库网</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">浙江大学管理学院市场营销学系教授、品质管理硕士学位项目论文导师王小毅和团队正是<span style="color: black;">发掘</span>了“深度强化学习”为数字营销<span style="color: black;">行业</span>革命带来的巨大启示和变化,对其进行了一系列深入<span style="color: black;">科研</span>,并设计了一种基于DRL的<strong style="color: blue;">个性化<span style="color: black;">目的</span>定位策略</strong>。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">日前,王小毅与团队在管理科学<span style="color: black;">行业</span>国际顶刊<strong style="color: blue;">《MANAGEMENT SCIENCE》</strong>上<span style="color: black;">发布</span>了<span style="color: black;">相关</span>数智营销的<span style="color: black;">研究</span>成果<strong style="color: blue;">"<span style="color: black;">Deep Reinforcement Learning for Sequential Targeting</span></strong>"(运用深度强化学习实现动态连续定位营销)。</p><img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOuhMYdRnBZWmnmwvEm6urN2S5TibG3GRsmXyz0ricH8LicvHaGZxEhxNIA/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><a style="color: black;">Deep Reinforcement Learning for Sequential Targeting.pdf</a></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">成果在线<span style="color: black;">发布</span>,点击链接查看详情</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">到底</span>什么是“深度强化学习”?</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">人工智能算法是<span style="color: black;">怎样</span>助力数字营销发展的?</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">研究</span>团队都做了<span style="color: black;">那些</span>尝试?</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">本期【<strong style="color: blue;">“<span style="color: black;">研究</span>”</strong>】,让<span style="color: black;">咱们</span><span style="color: black;">一块</span>走进王小毅团队的“AI营销世界”。</p><img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOMCYdf7qr1A834rX094UT3fcxbYacv9nuicuguSHJXdxEIUEelaXqSzA/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">学者简介</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">王小毅:浙江大学管理学院市场营销学教授、博导,<strong style="color: blue;">品质管理硕士学位项目论文导师,</strong>数字化转型与脑机智能营销交叉<span style="color: black;">科研</span>专家,<span style="color: black;">得到</span>中国<span style="color: black;">商场</span>联合会科技进步一等奖等省部级奖励多项。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">在《管理世界》、Management Science、MarketingScience、JournalofMarketingResearch、InformationSystemsResearch等期刊<span style="color: black;">发布</span>论文,谷歌学术H指数为20。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;"><span style="color: black;">她们</span><span style="color: black;">为么</span>聚焦<span style="color: black;">设备</span>学习来</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;"><span style="color: black;">科研</span>营销定位策略?</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">目的</span>市场定位营销(Target Marketing)<span style="color: black;">一般</span><span style="color: black;">指的是</span>企业识别各个<span style="color: black;">区别</span>的购买者群体,<span style="color: black;">选取</span>其中一个或几个<span style="color: black;">做为</span><span style="color: black;">目的</span>市场,运用适当的市场营销组合,集中力量为<span style="color: black;">目的</span>市场服务,满足<span style="color: black;">目的</span>市场的需要。然而,在数字化时代,企业的营销策略越来越依赖于与消费者的<strong style="color: blue;">高频次互动</strong>和<span style="color: black;">针对</span><strong style="color: blue;">营销策略的快速<span style="color: black;">调节</span></strong>。此时,经典营销理论的传统思路和手段<span style="color: black;">办法</span>面临着越来越多的挑战。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">传统的定位营销策略<span style="color: black;">常常</span>只<span style="color: black;">思虑</span>与顾客进行一次性“买卖”,较为依赖前期策划和庞大的<span style="color: black;">宣传</span>投放,忽略了时间<span style="color: black;">原因</span>对消费者<span style="color: black;">行径</span>的影响和促销活动的连续性,成本巨大且结果不确定。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">实质</span><span style="color: black;">状况</span>中,企业需要在多个周期内依次决定向谁发放优惠券、<span style="color: black;">选取</span>哪个促销活动以及决定两个活动之间的等待时长。与此<span style="color: black;">同期</span>,消费者的真实偏好<span style="color: black;">亦</span>会随着时间的推移而快速改变。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">因此呢</span>,不论是学术界和产业界都迫切需要一种随着<span style="color: black;">持续</span>变化的顾客<span style="color: black;">行径</span>而进行<span style="color: black;">调节</span>的<strong style="color: blue;">适应性定位营销策略</strong>。</p><img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrO3jRurHzjiaxPpRvVpXichgHafK7fFMJq0M6Yjulboun55HhxxMLkeM8w/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">企业<span style="color: black;">平常</span>的动态连续促销策略</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">做为</span>一种新兴的人工智能算法,“深度强化学习”(<span style="color: black;">Deep Reinforcement Learning,DRL</span>
    </p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">)算法<span style="color: black;">能够</span>在<span style="color: black;">无</span>人类监督的<span style="color: black;">状况</span>下<span style="color: black;">连续</span>“学习”<span style="color: black;">加强</span>性能,对<span style="color: black;">处理</span><span style="color: black;">以上</span>问题<span style="color: black;">拥有</span>巨大的<span style="color: black;">潜能</span>。</p>

    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">“深度强化学习(DRL)”是一种基于奖励的学习<span style="color: black;">办法</span>,它<span style="color: black;">能够</span><span style="color: black;">帮忙</span>定位营销策略实现个性化、强适应性,但它<span style="color: black;">亦</span>面临着<span style="color: black;">许多</span>挑战,例如<span style="color: black;">针对</span>时间影响的处理;<span style="color: black;">针对</span><span style="color: black;">繁杂</span>消费者<span style="color: black;">行径</span>维度的处理;<span style="color: black;">针对</span>策略效果的<span style="color: black;">评定</span>等。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">正是基于此背景,王小毅教授与团队提出了这种<span style="color: black;">创立</span>在DRL算法<span style="color: black;">基本</span>上的<strong style="color: blue;">个性化<span style="color: black;">目的</span>定位策略</strong>
    </p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">,并<span style="color: black;">运用</span>一种基于量化的不确定性学习启发式<span style="color: black;">办法</span>来使DRL适应<span style="color: black;">繁杂</span>的消费者<span style="color: black;">行径</span>维度。</p>
    <img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrO368D8aBVibMX2ia5Fh0FCnetuibXuVuKBcov9nRYEF5uVfQSIRDIgORWw/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">采用双流法( Two-Stream Computations)的竞争网络架构(Double-Dueling Network Architecture)</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;"><span style="color: black;">研究</span>团队都做出了<span style="color: black;">那些</span>尝试?</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">人工智能的“自主学习”和<span style="color: black;">持续</span>升级,是<span style="color: black;">经过</span><strong style="color: blue;">“深度学习”</strong>和<strong style="color: blue;">“强化学习”</strong>等技术来实现的。深度学习<span style="color: black;">拥有</span>较强的感知能力,<span style="color: black;">然则</span>缺乏<span style="color: black;">必定</span>的决策能力;而强化学习<span style="color: black;">拥有</span>决策能力,对感知问题束手无策。<span style="color: black;">因此呢</span>,将两者结合起来,<span style="color: black;">优良</span>互补,就能为<span style="color: black;">繁杂</span>系统的感知决策问题<span style="color: black;">供给</span>新的<span style="color: black;">处理</span>思路,这<span style="color: black;">便是</span><span style="color: black;">咱们</span>所说的<strong style="color: blue;">“深度强化学习(DRL)”,</strong><span style="color: black;">亦</span>是一种更接近人类思维方式的人工智能<span style="color: black;">办法</span>。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">近年来,“深度强化学习(DRL)”取得了巨大突破,这种人工算法<span style="color: black;">经过</span>在给定奖励或<span style="color: black;">处罚</span>反馈的<span style="color: black;">状况</span>下,让系统以<span style="color: black;">持续</span>试错的方式与营销环境进行连续交互,从而寻找到“最佳策略”,以完成学习的过程。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">王小毅教授团队正是基于DRL人工算法,在序贯性定位营销设置的场景下,设计了这项<strong style="color: blue;">个性化定位营销策略。</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">这项策略<span style="color: black;">首要</span><span style="color: black;">运用</span>连续价格促销来吸引顾客的即时<span style="color: black;">重视</span>力并锁定<span style="color: black;">她们</span>,并在每两次价格促销之间<span style="color: black;">供给</span>一个非促销期(即冷却期),并且随着时间的推移<span style="color: black;">逐步</span><span style="color: black;">增多</span>冷却期的长度,以便顾客<span style="color: black;">调节</span>其价格参考点。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">序贯性定位营销(Sequential Targeting),即连续地对消费者开展促销<span style="color: black;">行径</span>。</p>

    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">科研</span><span style="color: black;">显示</span>,采用深度强化学习<span style="color: black;">办法</span><span style="color: black;">能够</span><span style="color: black;">处理</span>当前实施连续定位营销策略所面临的三大挑战:</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">01</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">前瞻性</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">平衡<span style="color: black;">机构</span>当前收入和<span style="color: black;">将来</span>收入</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">02</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">边探索学习边<span style="color: black;">得到</span>市场<span style="color: black;">报答</span></strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">最大化利润,<span style="color: black;">同期</span><span style="color: black;">经过</span>探索<span style="color: black;">研发</span><span style="color: black;">持续</span>学习</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">03</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">可扩展性</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">应对高维状态和营销政策空间</p><img src="https://mmbiz.qpic.cn/mmbiz_jpg/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrO3iczjiafoEcmABq22ZQicgg9Mw4kF4EkcibuKzO3pwBzIuLz0NjkeaWNKQ/640?wx_fmt=jpeg&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">照片</span><span style="color: black;">源自</span>:©千库网</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">为了更好地使DRL能够适应<span style="color: black;">繁杂</span>的消费者<span style="color: black;">行径</span>维度,<span style="color: black;">科研</span>团队又提出了一种基于量化的<strong style="color: blue;">不确定性学习启发式算法</strong>,以实现<span style="color: black;">有效</span>的探索<span style="color: black;">研发</span>。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">经过</span><span style="color: black;">评定</span>,其结果<span style="color: black;">显示</span>,平均<span style="color: black;">状况</span>下,采用这种新算法代理(agent)产生的<span style="color: black;">长时间</span>收入比采用传统<span style="color: black;">办法</span>所产生的收入多26.75%,学习速度<span style="color: black;">亦</span>比所有基准中,其他产业界常用算法模型的速度快76.92%。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">为了更好地理解<span style="color: black;">科研</span>结果<span style="color: black;">背面</span>的潜在机制,<span style="color: black;">科研</span>团队进行了多种可解释性算法的<span style="color: black;">科研</span>,能够<span style="color: black;">诠释</span>个体和群体水平上学习最优策略的<span style="color: black;">行径</span>模式。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">另外</span>,王小毅教授与团队合作者提出的“模拟”在线测试环境为DRL训练和测试构建了用户<span style="color: black;">行径</span>模拟器,为平台<span style="color: black;">供给</span>了一种节省成本的方式来学习DRL代理,而无需在现实世界中运行<span style="color: black;">海量</span>、测试。</p><img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOWAJ4gs3sKEXCZVGicOFk4k6dBsM8biaRCgrKLwhkQQVxUMI5xY3ULBDA/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">照片</span><span style="color: black;">源自</span>:©千库网</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">然而,不得不承认的是人工智能算法在营销<span style="color: black;">行业</span>的应用仍然面临着<span style="color: black;">有些</span>挑战。<span style="color: black;">因为</span>算法依靠训练数据和计算资源,需要进行<span style="color: black;">海量</span>的实验和优化来<span style="color: black;">加强</span>算法的效率和准确性;<span style="color: black;">因为</span>消费者<span style="color: black;">行径</span>的<span style="color: black;">繁杂</span>性和不确定性,需要<span style="color: black;">更加多</span>的数据和模型来解释和预测消费者<span style="color: black;">行径</span>;<span style="color: black;">因为</span><span style="color: black;">运用</span>DRL算法的实时决策和<span style="color: black;">调节</span>需要在<span style="color: black;">短期</span>内做出,<span style="color: black;">因此呢</span>需要<span style="color: black;">创立</span>一个实时决策系统来支持DRL算法的应用。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">为<span style="color: black;">认识</span>决这些挑战,王小毅教授团队正在进一步<span style="color: black;">奋斗</span>开展<span style="color: black;">开发</span>工作,有望在今年<span style="color: black;">供给</span>一个更加通用和普适的<strong style="color: blue;">DRL框架</strong>,框架<span style="color: black;">亦</span><span style="color: black;">能够</span>很容易地推广到其他<span style="color: black;">目的</span>市场营销场景,例如基于位置的服务、在线流<span style="color: black;">媒介</span>或在线教育的序贯<span style="color: black;">举荐</span>,可供平台和营销人员在<span style="color: black;">平常</span>实践中<span style="color: black;">容易</span><span style="color: black;">运用</span>。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">该框架将基于量化的不确定性学习启发式<span style="color: black;">办法</span>,结合实时决策系统和用户<span style="color: black;">行径</span>模拟器,以<span style="color: black;">加强</span>DRL算法的效率和准确性,并<span style="color: black;">帮忙</span>企业更好地理解和预测消费者<span style="color: black;">行径</span>,制定更加智能和有效的营销策略。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">总体而言,王小毅教授团队的<span style="color: black;">科研</span>结果<span style="color: black;">表示</span>了“深度强化学习”<span style="color: black;">办法</span>在优化定位营销策略以最大化企业<span style="color: black;">长时间</span>收入方面的巨大<span style="color: black;">潜能</span>,证明了这种<span style="color: black;">办法</span>在数字营销<span style="color: black;">行业</span>能够产生颠覆性影响。</p><img src="https://mmbiz.qpic.cn/mmbiz_jpg/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOEzdYHmpVuOsLBkbE6dlIxC9m2nPh31ibspicN4rlicVbHBggPdY3oMZGQ/640?wx_fmt=jpeg&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">照片</span><span style="color: black;">源自</span>:©千库网</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><strong style="color: blue;"><span style="color: black;">她们</span>的<span style="color: black;">科研</span>对</strong></span></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><strong style="color: blue;">品牌价值和数字营销有何<span style="color: black;">道理</span>?</strong></span></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">基于“深度强化学习”的营销策略对<span style="color: black;">提高</span>数字化营销效率有何<span style="color: black;">功效</span>?对助力企业制定数字化营销策略有何<span style="color: black;">道理</span>?</p>

    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">其实,这项听起来“<span style="color: black;">奥秘</span>”的基于算法驱动的营销<span style="color: black;">科研</span>成果,与品牌、企业的息息<span style="color: black;">关联</span>,且在企业数字化转型中<span style="color: black;">拥有</span>重要价值与战略<span style="color: black;">道理</span>。<span style="color: black;">详细</span><span style="color: black;">来讲</span>,它<span style="color: black;">能够</span><span style="color: black;">帮忙</span>企业更加<span style="color: black;">精细</span>地进行市场定位和<span style="color: black;">目的</span>营销,<span style="color: black;">加强</span>数字化营销的效率和<span style="color: black;">精细</span>度,实现业务增长和品牌价值的<span style="color: black;">提高</span>。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">1</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">更好<span style="color: black;">认识</span>消费者的<span style="color: black;">需要</span>和偏好</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">“深度强化学习”<span style="color: black;">能够</span><span style="color: black;">帮忙</span>企业更好地<span style="color: black;">认识</span>消费者的<span style="color: black;">需要</span>和偏好,从而<span style="color: black;">精细</span>地进行市场定位和<span style="color: black;">目的</span>营销。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">经过</span>对消费者数据的深度分析,企业<span style="color: black;">能够</span><span style="color: black;">认识</span>消费者的<strong style="color: blue;"><span style="color: black;">行径</span>模式和心理特征</strong>,从而制定更加<span style="color: black;">精细</span>的营销策略。例如,企业<span style="color: black;">能够</span><span style="color: black;">按照</span>强化学习算法的分析结果,适时<span style="color: black;">调节</span><span style="color: black;">制品</span>定价、<span style="color: black;">制品</span><span style="color: black;">举荐</span>、<span style="color: black;">宣传</span>投放等策略,以<span style="color: black;">加强</span>消费者的消费忠诚度和购买转化率。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">2</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">优化营销<span style="color: black;">途径</span>,<span style="color: black;">提高</span>效率</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">“深度强化学习”<span style="color: black;">能够</span><span style="color: black;">帮忙</span>企业优化营销<span style="color: black;">途径</span>,<span style="color: black;">加强</span>数字化营销的效率和<span style="color: black;">精细</span>度。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">经过</span>对<span style="color: black;">区别</span>营销<span style="color: black;">途径</span>的数据进行分析和比较,企业<span style="color: black;">能够</span><span style="color: black;">认识</span><span style="color: black;">每一个</span><span style="color: black;">途径</span>的贡献度和效果,从而制定更加科学的<span style="color: black;">途径</span>策略。例如,企业<span style="color: black;">能够</span><span style="color: black;">按照</span>强化学习算法的分析结果,适时<span style="color: black;">调节</span>各个<span style="color: black;">途径</span>的投放比例和投放策略,以<span style="color: black;">加强</span>数字化营销的效率和<span style="color: black;">精细</span>度。</p>

    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">3</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">实现数字化营销的精细化和智能化</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">深度强化学习算法<span style="color: black;">能够</span><span style="color: black;">帮忙</span>企业实现数字化营销的精细化和智能化。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">经过</span>对消费者数据的深度学习和分析,企业<span style="color: black;">能够</span><span style="color: black;">认识</span>消费者的<span style="color: black;">行径</span>和偏好,从而实现个性化的营销和服务。例如,企业<span style="color: black;">能够</span><span style="color: black;">按照</span>强化学习算法的分析结果,针对<span style="color: black;">区别</span>的消费者群体,<span style="color: black;">供给</span><span style="color: black;">区别</span>的营销<span style="color: black;">方法</span>和服务,以<span style="color: black;">加强</span>消费者的满意度和忠诚度。</p><img src="https://mmbiz.qpic.cn/mmbiz_jpg/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOO92fxXf0h8m20xf0m0A52jnOtrbAZeMKWskiazB9ib6VKlOdicHSWnrww/640?wx_fmt=jpeg&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">图片<span style="color: black;">源自</span>:©千库网</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">基于这项<span style="color: black;">科研</span>成果,王小毅团队与阿里巴巴合作改进数字营销经营<span style="color: black;">办法</span>论、与浙江中烟合作设计智慧营销大脑,为传统零售<span style="color: black;">商场</span>的制定自有品牌策略等,为企业数智创新<span style="color: black;">供给</span>了新的路径和解题方式。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">日前,王小毅教授联合阿里妈妈发布的DEEPLINK模型,<span style="color: black;">仔细</span>解析了消费者从种草,到首购,在到复购的消费路径。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">如果以“个性化定位营销策略”来分析<span style="color: black;">诠释</span>模型中看似<span style="color: black;">繁杂</span>无序的消费路径,<span style="color: black;">能够</span>将其拆解为:Discover<span style="color: black;">发掘</span>—Engage种草—Enthuse热爱—Perform行动—Initial首购—Numerous复购—Keen忠诚。该模型是基于过去围绕<span style="color: black;">营销</span>漏斗的AIPL模型(Awareness认知-Interest兴趣-Purchase购买-Loyalty忠实)的进一步升级。</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">这般</span>的路径分析能够<span style="color: black;">帮忙</span>企业更好地理解和预测消费者<span style="color: black;">行径</span>,制定更加智能和有效的营销策略。</p><img src="https://mmbiz.qpic.cn/mmbiz_png/vk4jak4unGonVd9GCgmA8CUJ5dhFbWrOIYxjg5TUCr674kKibnslC5ickN03RB4bLYTcQzjE40IWf8GD2lfW9Zog/640?wx_fmt=png&amp;wxfrom=5&amp;wx_lazy=1&amp;wx_co=1&amp;tp=webp" style="width: 50%; margin-bottom: 20px;">
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><strong style="color: blue;">附:论文摘要</strong></p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">Deep reinforcement learning (DRL) has opened up many unprecedented opportu-nities in revolutionizing the digital marketing field. In this study, we designed a DRL-based personalized targeting strategy in a sequential setting. We show that the strategy is able to address three important challenges of sequential targeting: (1) forward looking (balancing between a firm’s current revenue and future revenues), (2) earning while learning (maximizing profits while continuously learning through exploration-exploitation), and (3) scalability (cop-ing with a high-dimensional state and policy space). We illustrate this through a novel design of a DRL-based artificial intelligence (AI) agent. To better adapt DRL to complex consumer behavior dimensions, we proposed a quantization-based uncertainty learning heuristic for effi-cient exploration-exploitation. Our policy evaluation results through simulation suggest that the proposed DRL agent generates 26.75% more long-term revenues than can the non-DRL approaches on average and learns 76.92% faster than the second fastest model among all bench-marks. Further, in order to better understand the potential underlying mechanisms, we con-ducted multiple interpretability analyses to explain the patterns of learned optimal policy at both the individual and population levels. Our findings provide important managerial- relevant and theory-consistent insights. For instance, consecutive price promotions at the begin-ning can capture price-sensitive consumers’ immediate attention, whereas carefully spaced nonpromotional “cooldown” periods between price promotions can allow consumers to adjust their reference points. Additionally, consideration of future revenues is necessary from a long- term horizon, but weighing the future too much can also dampen revenues. In addition, analy-ses of heterogeneous treatment effects suggest that the optimal promotion sequence pattern highly varies across the consumer engagement stages. Overall, our study results demonstrate&nbsp;DRL’s potential to optimize these strategies’ combination to maximize long-term revenues.</p>
    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;">本期素材 | ZJUSOM官微</p>

    <p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">本期责编 | ZUPUC品牌推广</span></p>




lanbingkj 发表于 2024-9-9 08:02:36

感谢你的精彩评论,为我的思绪打开了新的窗口。

m5k1umn 发表于 2024-10-4 01:08:39

这篇文章真的让我受益匪浅,外链发布感谢分享!

m5k1umn 发表于 2024-10-4 19:28:42

“沙发”(SF,第一个回帖的人)‌

4zhvml8 发表于 2024-10-5 06:45:44

i免费外链发布平台 http://www.fok120.com/

4lqedz 发表于 2024-10-11 03:49:53

“板凳”(第三个回帖的人)‌
页: [1]
查看完整版本: 他的UTD科研证明,AI算法能够给数字营销带来“巨变”……| 学术对话No.1