概述
Last updated
Last updated
大数据时代，利用人工智能等前沿技术进行改造升级已成为各行各业的必然趋势。其中，利用以个人为中心的用户数据建立千人千面的智能服务系统，已广泛应用于多个领域，成为学术界与工业界长久以来的研究热点与研究重点。而其中游戏产业为用户智能服务研究提供了良好的数据基础和丰富的研究问题，同时也带来了独特的挑战。
首先，游戏中的用户数据具有类型丰富、日志齐全、场景多样等特点，如何从多源异构的游戏数据中构建高效的用户动态表征是用户智能服务研究的数据基础和关键难点。例如，用户个人信息能够总体描述用户基本特征，用户行为日志能够精细刻画用户动作历史，用户关系图谱能够多角度呈现用户与其他对象之间的交互联系，用户多模态数据如图片、轨迹、键鼠、聊天、语音等能够反映用户实时动态
其次，游戏受众各不相同，玩家群体呈现泛化特征，而玩家个体存在个性偏好，如何结合用户表征，理解用户潜在特性和行为因果模式，提出具有可解释性的用户建模算法是用户智能服务研究的核心问题和模型基础。一方面，玩家按照玩法偏好可以划分为不同群体，如“风景党”、“任务狂”、“对战流”等，玩家基于社交关系或者游戏设计会加入临时团体如队伍或者固定团体如帮会，队伍或者帮会会呈现出整体趋同的特点；另一方面，每一名玩家性格、偏好迥然不同，平均每天都会有上万条游戏行为事件，在游戏世界中留下独一无二的活动轨迹。。
最后也是最重要的一点，游戏世界自由度不断提升，玩法元素层出不穷，衍生出极其丰富的应用场景，如何基于用户动态表征和用户建模算法，结合特定业务领域知识，转化出可行的用户服务应用成为用户智能服务研究的重要目标和落地难点。从玩家角度，道具/礼包推荐、侠侣/师徒推荐、装备/时装推荐等个性化策略帮助个体玩家快速成长和更好地享受游戏快乐；从运营角度，对战匹配、外挂检测、活动模拟等整体配置促进游戏整体公平性、可玩性和可持续性；从程序角度，资源预加载、延迟预测、卡顿检测等客户端或服务端优化协助开发人员更好维护玩家游戏体验。
网易游戏伏羲实验室用户画像组(以下简称为Fuxi-UP，https://fuxi.163.com/tech.html#yhhx)依托游戏平台海量数据，形成了游戏场景下的标准化大数据集，以“背景-定义-挑战-评估-数据”的形式提炼出一系列的开放研究课题，致力于面向海内外青年学者搭建游戏用户研究产学研合作平台。Fuxi-UP已建立起以“用户表征-用户建模-服务应用”为核心的研究体系（如图1所示），与多款在营游戏合作，为千万级玩家提供贴心服务，同时与浙江大学、清华大学、中国科学技术大学、天津大学、中科院计算所等国内知名高校和潘纲、崔鹏、陈刚、刘淇、况琨、赵洪科、赵莎、陈龙彪、毕经平、姚迪、赵翔宇等优秀学者合作，在KDD、WWW、ICDE、AAAI、MM、ICDM、CIKM、IEEE TC、IEEE TKDE、ACM TKDD、ACM TOIS、IEEE TG等人工智能领域顶级会议和期刊上发表多篇学术论文，并在国际游戏领域顶级学术会议IEEE CoG 2020荣获最佳论文奖。
针对用户表征，Fuxi-UP应用人工智能技术帮助搭建游戏用户标签体系，学习用户动态行为行为嵌入和用户动态网络嵌入，捕获游戏玩家自身和玩家之间的深层次动态关联；针对方法建模，Fuxi-UP结合人工智能技术，对关系网络中的用户行为事件建模，通过模型可解释、因果分析、联邦学习、自动机器学习、在线学习等先进技术方法，有效捕获用户行为背后的隐藏规律和隐含状态；同时在领域知识加持下，Fuxi-UP运用人工智能技术能加速反外挂、用户流失/付费预测、玩家胜率预测、项目价值预估/预测等任务，同时针对用户个人提供定制化的推荐、匹配等服务，高效孵化出优质智能服务应用。
当前已成型的开放研究课题包括：（不断更新中）
当前已成型的开放工程课题包括：（不断更新中）
同时为帮助领域外学者快速了解游戏场景和开展研究，整理相关游戏背景知识和游戏数据说明如下：
Fuxi-UP组论文兴趣既包括针对通用AI问题的研究性论文，也包括结合业务落地的工业性论文，具体的论文流程请参考用户画像组论文流程。
Fuxi-UP组期待与高校学术力量进行深入合作，合作形式包括且不限于学术顾问、横向合作、论文合作以及联合申办学术活动等。
具体合作申请事宜请参考高校合作申请流程。
如有疑问，欢迎邮件联系咨询：
Email: hztaojianrong@corp.netease.com, wurunze1@corp.netease.com
Fuxi-UP组为学术界提供来自工业界的实际问题及数据，举办了相关的数据竞赛， 欢迎来自学界和业界的同仁积极参与：
Minghao Zhao, Le Wu, Yile Liang, Lei Chen, Jian Zhang, Qilin Deng, Kai Wang, Xudong Shen, Tangjie Lyu and Runze Wu, Investigating Accuracy-Novelty Performance for Graph-based Collaborative Filtering, SIGIR'2022 (accepted)
Xianyang Qi, Jiashu Pu, Shiwei Zhao, Jianrong Tao and Runze Wu, A GNN-Enhanced Game Bot Detection Model for MMORPGs, PAKDD'2022 (accepted)
Wenbin Li, Xiaokai Chu, Yueyang Su, Di Yao, Shiwei Zhao, Runze Wu, Shize Zhang, Jianrong Tao, Hao Deng and Jingping Bi, FingFormer: Contrastive Graph-based Finger Operation Tranformer for Unsupervised Mobile Game Bot Detection, WWW'2022 (accepted)
Jiashu Pu, Jianshi Lin, Xiaoxi Mao, Xudong Shen, Jianrong Tao, Yue Shang and Runze Wu, Unsupervised representation learning of player behavioral data with Confidence Guided Masking, WWW'2022 (accepted)
Chuang Zhao, Hongke Zhao, Yong Ge, Runze Wu and Xudong Shen, Winning Tracker: A New Model for Real-time Winning Prediction in MOBA Games, WWW'2022 (accepted)
Jian Zhang, Minghao Zhao, Runze Wu and Qi Xuan, Center-Oriented Attentive Temporal Pooling for Link Prediction on Dynamic Networks, AAAI'2022 Workshop (accepted)
Lei Zhang, Xiang Wang, Chuang Zhao, Hongke Zhao, Rui Li and Runze Wu, Co-Promotion Predictions of Financing Market and Sales Market: A Cooperative-Competitive Attention Approach, AAAI'2022 (accepted)
Chuang Zhao, Hongke Zhao, Runze Wu, Qiling Deng, Yu Ding, Jianrong Tao and Changjie Fan, Multi-dimensional Prediction of Guild Health in Online Games: A Stability-AwareMulti-task Learning Approach, AAAI'2022 (accepted)
Sha Zhao, Junwei Fang, Shiwei Zhao, Runze Wu, Jianrong Tao, Shijian Li and Gang Pan, T-Detector: A Trajectory based Pre-trained Model for Game Bot Detection in MMORPGs, ICDE'2022 (accepted)
Manhu Qu, Jie Huang, Hao Deng, Runze Wu, Xudong Shen, Jianrong Tao and Tangjie Lyu, EasySM: A data-driven intelligent decision support system for server merge, AAAI'2022 Demo (accepted)
Jie Huang, Qi Liu, Fei Wang, Zhenya Huang, Songtao Fang, Runze Wu, Enhong Chen, Yu Su and Shijin Wang, Group-Level Cognitive Diagnosis: A Multi-Task Learning Perspective, ICDM'2021 (accepted)
Qiling Deng, Kai Wang, Minghao Zhao, Runze Wu, Yu Ding, Zhene Zou, Yue Shang, Jianrong Tao and Changjie Fan, Build Your Own Bundle - A Neural Combinatorial Optimization Method, ACMMM'2021 (accepted)
Qiling Deng & Hao Li, Kai Wang, Zhipeng Hu, Runze Wu, Linxia Gong, Jianrong Tao, Changjie Fan and Peng Cui, "Globally Optimized Matchmaking in Online Games", KDD' 2021 (accepted)
Yin Gu, Qi Liu, Kai Zhang, Zhenya Huang, Runze Wu, Jianrong Tao, "NeuralAC: Learning Cooperation and Competition Effects for Match Outcome Prediction", AAAI' 2021 PDF
Kai Wang, Zhene Zou, Qilin Deng, Jianrong Tao, Runze Wu, Changjie Fan, Liang Chen, Peng Cui, "Reinforcement Learning with a Disentangled Universal Value Function for Item Recommendation", AAAI' 2021 PDF
Jianrong Tao, Yu Xiong, Shiwei Zhao, Yuhong Xu, Jianshi Lin, Runze Wu, Changjie Fan, "XAI-Driven Explainable Multi-view Game Cheating Detection", CoG' 2020 (Best Paper) PDF
Shiwei Zhao, Runze Wu, Jianrong Tao, Manhu Qu, Hao Li, Changjie Fan, "Multi-source Data Multi-task Learning for Profiling Players in Online Games", CoG' 2020 PDF
Linxia Gong, Xiaochuan Feng, Dezhi Ye, Hao Li, Runze Wu, Jianrong Tao, Changjie Fan, Peng Cui, "OptMatch: Optimized Matchmaking via Modeling the High-Order Interactions on the Arena", KDD' 2020 PDF
Qilin Deng, Kai Wang, Minghao Zhao, Zhene Zou, Runze Wu, Jianrong Tao, Changjie Fan, Liang Chen, "Personalized Bundle Recommendation in Online Games", CIKM' 2020 PDF
Runze Wu, Hao Deng, Jianrong Tao, Changjie Fan, Qi Liu, Liang Chen, "Deep Behavior Tracing with Multi-level Temporality Preserved Embedding", CIKM' 2020 PDF
Kai Wang, Hao Li, Linxia Gong, Jianrong Tao, Runze Wu, Changjie Fan, Liang Chen, Peng Cui, "Match Tracing: A Unified Framework for Real-time Win Prediction and Quantifiable Performance Evaluation", CIKM' 2020 PDF
Angyu Zheng, Liang Chen, Fenfang Xie, Jianrong Tao, Changjie Fan, Zibin Zheng, "Keep You from Leaving: Churn Prediction in Online Games", DASFAA' 2020 PDF
Jianrong Tao, Linxia Gong, Changjie Fan, Longbiao Chen, Dezhi Ye, Sha Zhao, "GMTL: A GART Based Multi-task Learning Model forMulti-Social-Temporal Prediction in Online Games", CIKM' 2019 PDF
Jianrong Tao, Jianshi Lin, Shize Zhang, Sha Zhao, Runze Wu, Changjie Fan, Peng Cui, "MVAN: Multi-view Attention Networks for Real Money Trading Detection in Online Games", KDD' 2019 PDF
Jianrong Tao, Jiarong Xu, Linxia Gong, Yifu Li, Changjie Fan, Zhou Zhao，“NGUARD: A Game Bot Detection Framework for NetEase MMORPGs”， KDD' 2018 PDF
Shiwei Zhao and Runze Wu, Jianrong Tao, Manhu Qu, Minghao Zhao and Changjie Fan, perCLTV: A General System for Personalized Customer Lifetime Value Prediction in Online Games, ACM TOIS (accepted)
Xiang Wen & Shiwei Zhao, Haobo Wang, Runze Wu, Manhu Qu, Tianlei Hu, Gang Chen, Jianrong Tao and Changjie Fan, Multi-Source Multi-Label Learning for User Profiling in Online Games, IEEE TMM (accepted)
Anpeng Wu, Junkun Yuan, Kun Kuang, Bo Li, Runze Wu, Qiang Zhu, Yueting Zhuang and Fei Wu, Learning Decomposed Representations for Treatment Effect Estimation, IEEE TKDE (accepted)
Chongrong Fang, Haoyu Liu, Mao Miao, Jie Ye, Lei Wang, Wansheng Zhang, Daxiang Kang, Biao Lyu, Shunmin Zhu, Peng Cheng and Jiming Chen, Towards Automatic Root Cause Diagnosis of Persistent Packet Loss in Cloud Overlay Network, IEEE/ACM TON (accepted)
Ge Fan, Biao Geng, Jianrong Tao, Kai Wang and Changjie Fan, PPPNE: Personalized Proximity Preserved Network Embedding, NEUROCOMPUTING (accepted)
Qiling Deng, Minghao Zhao, Kai Wang, Runze Wu, Xudong Shen, Jianrong Tao and Changjie Fan, "Find Your Organization in MMORPGs", IEEE Trans. on Games (accepted)
Kun Kuang, Hengtao Zhang, Runze Wu, Fei Wu, Yueting Zhang and Aijun Zhang, "Balance-Subsampled Stable Prediction across Unknown Test Data", ACM TKDD (accepted)
Minghao Zhao, Qilin Deng, Kai Wang, Runze Wu, Jianrong Tao, Changjie Fan, Liang Chen and Peng Cui, "Bilateral Filtering Graph Convolutional Network for Social Recommendation in Power-Law Networks", ACM TOIS (accepted)
Jiarong Xu and Yifan Luo, Jianrong Tao and Changjie Fan, Zhou Zhao and Jiangang Lu, "NGUARD+: An Attention-based Game Bot Detection Framework via Player Behavior Sequences", ACM TKDD PDF
Kun Kuang, Peng Cui, Hao Zou, Bo Li, Jianrong Tao, Fei Wu, Shiqiang Yang, "Data-Driven Variable Decomposition for Treatment Effect Estimation", IEEE TKDE PDF
