姓名:兴军亮
职称:研究员
邮件:jlxing@tsinghua.edu.cn
个人主页:https://pi.cs.tsinghua.edu.cn/people/#faculty
教育经历:
2003.09-2007.07 工学理学双学士 西安交通大学计算机科学与技术系/数学与应用数学系
2007.09-2012.07 工学博士 清华大学计算机科学与技术系
工作经历:
2012.07-2015.10 中科院自动化所 模式识别国家重点实验室 助理研究员
- 2012.12-2013.11 新加坡国立大学 访问研究员
- 2015.04-2015.10 微软亚洲研究院 “铸星计划”访问教授
2015.11-2018.10 中科院自动化所 模式识别国家重点实验室 副研究员
2018.11-2022.01 中科院自动化所 模式识别国家重点实验室 研究员
2022.02至今 清华大学计算机科学与技术系 研究员
研究方向:计算机视觉、计算机博弈、人机交互学习
研究概况:我的研究工作主要围绕智能感知与决策方向展开:在智能感知方向,早期研究主要关注无约束条件下的视觉目标检测、配准、跟踪、分割、识别等任务,近年来开始尝试融合多模态信息进行复杂态势信息感知与分析;在智能决策方向,针对复杂系统决策问题存在的难建模、难计算、难解释挑战,通过结合传统博弈理论方法对问题进行表示建模、最新机器学习算法技术对模型进行近似求解,以及人机交互过程对模型进行分析解释,逐渐形成了博弈交互学习的研究路线。相关研究工作在CCF-A类国际会议和刊物上发表论文70余篇、谷歌学术引用超过17000次,其中人像属性关系感知算法获CCF-A类国际会议ACM Multimedia 2013最佳论文奖、无锚点单阶段在线多物体跟踪算法获计算机视觉领域重要国内会议PRCV 2020最佳论文奖、大规模不完美信息博弈算法AlphaHoldem获CCF-A类国际会议AAAI 2022卓越论文奖。研发的智能感知与决策算法十余次在国内外重要技术竞赛中获奖,并多次在重要实际应用场景中得到验证和大规模落地应用
讲授课程:
《计算机视觉》:清华大学,研究生课程,48学时,2024年春季学期至今
《程序设计基础》:清华大学,本科生课程,48学时,2023年秋季学期至今
《博弈论》:中国科学院大学,本科生课程,20学时,2024年春季学期
《计算博弈原理与应用》:中国科学院大学,研究生课程,40学时,2019年秋季学期-2023年秋季学期
《博弈论》:中国科学院大学,研究生课程,20学时,2017年春季学期~2019年春季学期
科研项目:
• 视觉态势感知和博弈决策:国家自然科学基金委优秀青年基金项目,200万,负责人,在研
• 大规模不完美信息博弈高效求解方法研究:国家自然科学基金委面上项目,59万,负责人,在研
• 对抗态势感知基础理论和关键技术研究:科技部“科技创新2030-新一代人工智能”重大项目子课题,593万,负责人,结题
• 无约束人脸识别问题研究:国家自然科学基金委面上项目,63万,负责人,结题
• 基于区分性模型学习的综合在线多物体检测、跟踪和分割:国家自然科学基金委青年基金项目,25万,负责人,结题
• 人脸和人体视觉算法相关研究与应用:华为公司企业横向合作系列项目(2016~2020年),~500万,负责人,结题
研究介绍:
我近些年来关于智能博弈决策的一些内容请观看如下一些公开视频讲座:
兴军亮:智能博弈学习技术与应用 |【自动化所2020年“人工智能”暑期学校】哔哩哔哩_bilibili
兴军亮:重新定义电子游戏,人工智能最好的技术试验场?|【格致论道】哔哩哔哩_bilibili
兴军亮:大规模不完美信息博弈平台与算法 |【RLChina 前沿讲习班】哔哩哔哩_bilibili
研究成果:
代表性期刊论文(* indicates the corresponding author, # indicates a co-first author):
[1] Junliang Xing, Zhiheng Niu, Junshi Huang, Weiming Hu, Xi Zhou, and Shuicheng Yan. Towards Robust and Accurate Multi-view and Partially-occluded Face Alignment. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), vol. 40, no. 4, pp. 987-1001, 2018.
[2] Junliang Xing, Weiming Hu, Haizhou Ai, and Shuicheng Yan. FatRegion: A Fast, Adaptive, Tree-structured Region Extraction Approach. IEEE Transactions on Circuits, System and Video Technology (TCSVT), vol. 28, no. 3, pp. 601-615, 2018.
[3] Junliang Xing, Kai Li, Weiming Hu, Chunfeng Yuan, and Haibin Ling. Diagnosing Deep Learning Models for High Accuracy Age Estimation from A Single Image. Pattern Recognition (PR), vol. 66, pp. 95-105, 2017.
[4] Junliang Xing, Haizhou Ai, Liwei Liu, and Shihong Lao, Multiple Players Tracking in Sports Video: A Dual-Mode Two-Way Bayesian Inference Approach with Progressive Observation Modeling, IEEE Transactions on Image Processing (TIP), vol. 20, no. 6, pp. 1652-67, 2011.
[5] Wenhan Luo, Junliang Xing*, Anton Milan, Xiaoqin Zhang, Wei Liu, Tae-Kyun Kim. Multiple Object Tracking: A Literature Review. Artificial Intelligence (AI), vol. 293, pp. 1–23, 2021.
[6] Xiaoshu Shao#, Junliang Xing#, Jiangjing Lv, Xiangdong Zhou, Yu Shi, Steve Maybank. Robust Face Alignment via Deep Progressive Reinitialization and Adaptive Error-driven Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 10.1109/TPAMI.2021.3073593, 2021.
[7] Jian Zhao, Junliang Xing*, Lin Xiong, Shuicheng Yan, Jiashi Feng. Recognizing Profile Faces by Imagining Frontal View. International Journal of Computer Vision (IJCV), vol. 128, pp. 460–478, 2020.
代表性会议论文(* indicates the corresponding author, # indicates a co-first author):
[1] Junliang Xing, Zhiheng Niu, Junshi Huang, Weiming Hu, Shuicheng Yan. Towards Multi-view and Partially-occluded Face Alignment. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014.
[2] Junliang Xing, Jin Gao, Bin Li, Weiming Hu, Shuicheng Yan. Robust Object Tracking with Online Multi-lifespan Dictionary Learning. In IEEE International Conference on Computer Vision (ICCV), 2013.
[3] Junliang Xing, Haizhou Ai and Shihong Lao, Multi-Object Tracking through Occlusions by Local Tracklets Filtering and Global Tracklets Association with Detection Responses. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2009.
[4] Enmin Zhao, Renye Yan, Jinqiu Li, Kai Li, and Junliang Xing*. AlphaHoldem: High-Performance Artificial Intelligence for Heads-Up No-Limit Poker from End-to-End Reinforcement Learning. In AAAI Annual Conference on Artificial Intelligence (AAAI), 2022.(卓越论文奖)
[5] Hang Xu, Kai Li, Haobo Fu, Qiang Fu, and Junliang Xing*. AutoCFR: Learning to Design Counterfactual Regret Minimization Algorithms. In AAAI Annual Conference on Artificial Intelligence (AAAI), 2022.
[6] Yongxin Kang, Enmin Zhao, Kai Li, and Junliang Xing*. Exploration via State Influence Modeling. In AAAI Annual Conference on Artificial Intelligence (AAAI), 2021.
[7] Peixi Peng#, Junliang Xing#*, and Lili Cao. Hybrid Learning for Multi-agent Cooperation with Sub-optimal Demonstrations. In International Joint Conferences on Artificial Intelligence (IJCAI), 2020.
[8] Enmin Zhao, Shihong Deng, Yifan Zang, Yongxin Kang, and Junliang Xing*. Potential Driven Reinforcement Learning for Hard Exploration Tasks. In International Joint Conferences on Artificial Intelligence (IJCAI), 2020.
[9] Zongwei Zhou, Yangxi Li, Jin Gao, Junliang Xing, Liang Li, and Weiming Hu. Anchor-Free One-Stage Online Multi-object Tracking. In Chinese Conference on Pattern Recognition and Computer Vision (PRCV), 2020.(最佳论文奖)
[10] Luoqi Liu, Hui Xu, Junliang Xing, Si Liu, Xi Zhou, Shuicheng Yan. “Wow! You Are So Beautiful Today!” In ACM International Conference on Multimedia (ACM MM), 2013.(最佳论文奖)