Activities

2015年10月26日至30日，第23届ACM Multimedia会议在澳大利亚布里斯班举行举行。PALM实验室的周颖同学参加了此次会议，并且代表实验室做了墙展交流。

会议开幕式

周颖及其墙展海报

题目： 场景文字理解的若干关键技术研究

讲者： 白翔 教授，华中科技大学

摘要： 随着智能终端的普及和移动互联网的飞速发展，场景文字识别（Photo OCR）越来越引起学术界和工业界的重视。图片中的文字包含丰富的高层语义，因此如何去检测和识别自然场景中的文字具有广泛的技术应用前景。在这次报告中，我将首先介绍场景文字检测和识别的基本方法、前沿技术以及相关应用。然后，我将介绍我们在此领域的最新研究成果，包括：基于字符定位的场景文字识别算法(CVPR14)；基于对称性的文字区域检测算法(CVPR15); 场景文字语种识别方法（ICDAR15）；及图像序列识别神经网络等。

个人简介：白翔博士现为华中科技大学电子信息与通信学院教授，博士生导师，担任国家防伪工程中心副主任。他的研究领域为计算机视觉与模式识别，具体包括目标识别、形状分析、自然场景文字识别及智能系统。他已在计算机视觉与模式识别相关的国际权威期刊或顶级会议发表论文30余篇。他的研究工作曾获得微软学者2007，首届国家自然科学基金优秀青年基金的资助。他的博士论文获得2012年全国优秀博士论文提名。担任中国图象图形学学会图象视频处理与通信专业委员会秘书长，人工智能学会模式识别专委会委员，计算机学会计算机视觉专业组委员，计算机学会人工智能与模式识别专业委员会通讯委员，视觉与学习青年研讨会（VALSE）在线委员。担任包括权威杂志PAMI、IJCV、TIP、TSMC、TNN、TMM、PR、CVIU、PRL、IVC等评审和顶级会议CVPR ,ICCV, NIPS, ECCV等TPC。

题目：大规模分布式内容搜索技术研究

讲者：陈汉华 教授，华中科技大学

摘要：近年来，各类分布式应用系统蓬勃发展，其发展呈现组织规模化、资源异构化、应用复杂化的特点。如何提升分布式内容搜索效率，成为制约分布式系统和应用发展的主要瓶颈之一。大规模分布式内容搜索主要面临大规模数据划分，分布式资源组织和复杂化索引构建三个方面的挑战。本研究围绕大规模分布式内容搜索的核心问题，结合对等网络、传感器网络、社交网络等分布式应用系统背景，探讨相关支撑技术。

个人简介：2010年从华中科技大学计机学院获得博士学位。2011年晋升为副教授，2014年晋升为教授。目前主持全国百篇优秀博士学位论文作者专项、国家自然科学基金优秀青年科学基金、国家自然科学基金面上项目、教育部-中移动科研基金项目、CCF-Intel青年提升计划项目等多个项目。在WWW、RTSS、ICNP、INFOCOM、ICDE、HPDC、IWQoS、IPDPS、ICPP、TC、TPDS、TKDE、TMC、TSC等国际会议和期刊上发表论文60余篇。获得国家科技进步二等奖1项（排名第四），湖北省科技进步二等奖1项。

Speaker:

Prof. Chengqi Zhang
Director, Centre for Quantum Computation & Intelligent Systems (QCIS)
University of Technology, Sydney
Australia

Title: Graph Processing and Mining in the Era of Big Data

Abstract

With the emergence and rapid proliferation of applications that deal with big graphs, such as web graphs (Google, Yahoo), social networks (Facebook, Twitter), e-commerce networks (Amazon, Ebay), and road networks, graph processing and mining has become increasingly prevalent and important in recent years. However, in the era of big data, the explosion and profusion of available graph data in a wide range of application domains rise up new challenges and opportunities in graph processing and mining.
Graph processing and mining is one of the research strengths in the centre for Quantum Computation and Intelligent Systems (QCIS) at the University of Technology, Sydney (UTS). In this talk, I will first investigate the new challenges for graph processing and mining in the era of big data. To tackle these challenges, I will introduce the recent research developments in QCIS in terms of new graph query semantics, new graph mining tasks, new query processing algorithms, new graph indexing techniques, and new computing paradigms. Finally, I will show our current achievements in building a general-purpose graph processing and mining system in QCIS centre, and discuss our potential future research directions.

Biography

Chengqi Zhang has been appointed as a Research Professor of Information Technology at The University of Technology, Sydney (UTS) since December 2001. He has been the Director of the UTS Research Centre for Quantum Computation & Intelligent Systems (QCIS) since April 2008. Chengqi Zhang obtained his PhD degree from the University of Queensland in 1991, followed by a Doctor of Science (DSc – Higher Doctorate) from Deakin University in 2002, all from computer science. He had been appointed by University of New England (UNE) from 1990 to 1998 as Lecturer, Senior Lecturer, and Associate Professor, then Deakin University from 1999 to 2001 as Associate Professor, then UTS from 2002 till now as Research Professor.
Prof. Zhang’s key areas of research are Distributed Artificial Intelligence, Data Mining and its applications. He has published more than 200 refereed research papers, including a number of papers in the first-class international journals, such as Artificial Intelligence, IEEE and ACM Transactions. He has delivered 14 keynote/invited speeches at international conferences over the last eight years. He has attracted 12 ARC grants of $4.7M. He has supervised 30+ PhD students in completion. He received NSW State Science and Engineering Award in Engineer and ICT category in 2011 and also UTS Chancellor research excellence award in Research Leadership category in 2011.
Prof. Zhang is a Fellow of the Australian Computer Society (ACS) and a Senior Member of the IEEE Computer Society (IEEE). He had been serving ARC as an ARC College of Expert from 2012 to 2014. He has been the Chair of the Australian Computer Science National Committee on Artificial Intelligence from 2005 till now. He was General Co-Chair of PAKDD 2014, WI/IAT 2018, and ICDM 2010. He is the General Co-Chair of KDD 2015 and he is also Local Arrangements Chair of IJCAI 2017.

2015年8月21-23日，中国计算机学会人工智能会议（CCFAI’15）在山西太原举行。PALM实验室教师耿新、张敏灵、薛晖、刘胥影参加了本次会议，耿新老师、张敏灵老师担任分组报告主持人。

2015年8月16-18日，第十五届中国机器学习会议（CCML’15）在四川成都举行。PALM实验室教师耿新、张敏灵、薛晖、刘胥影以及研究生李森、于菲参加了本次会议。耿新老师、张敏灵老师担任分组报告主持人，李森、于菲分别就论文《一种基于不定核的大间隔聚类算法》和《PL-forest:一种基于决策树集成的偏标记学习算法》做了口头报告。

2015年7月25-31日，第二十四届国际人工智能联合大会（IJCAI’15）在阿根廷布宜诺斯艾利斯举行。PALM实验室教师耿新、张敏灵参加了本次会议，并先后担任三场分组报告的Session Chair。耿新老师就论文《Pre-release prediction of crowd opinion on movies by label distribution learning》、张敏灵老师就论文《Solving the partial label learning problem: An instance-based approach》与《Towards class-imbalance aware multi-label learning》做了口头报告。

2015年5月8日至10日，VALSE2015会议在成都电子科技大学（沙城校区）举行。PALM实验室的耿新老师及侯鹏、邢超和任意同学参加了此次会议。耿老师担任了此次大会部分主持工作，侯鹏的海报《Pre-release Prediction of Crowd Opinion on Movies》参加了墙展交流。

耿老师为大会主持人。

侯鹏介绍他的post。

墙展交流。

In the recent IJCAI'15, three papers written by PALM members were accepted:

[1] Xin Geng and Peng Hou. Pre-release Prediction of Crowd Opinion on Movies by Label Distribution Learning. In: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI'15), Buenos Aires, Argentina, 2015, in press.

[2] Min-Ling Zhang, Fei Yu. Solving the Partial Label Learning Problem: An Instance-based Approach, In: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI'15), Buenos Aires, Argentina, 2015, in press.

[3] Min-Ling Zhang, Yu-Kun Li, Xu-Ying Liu. Towards Class-imbalance Aware Multi-label Learning, In: Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI'15), Buenos Aires, Argentina, 2015, in press.

Speaker:

Prof Jingdong Wang,
Visual Computing Group
Microsoft Research Asia

Title: Big Media Data: Search and Management

Abstract

The explosion of images, videos and other media data in the Internet, mobile devices, and desktops has attracted more and more interest in the Big Media research area. Big media opens great unprecedented opportunities to address many challenging computing problems. In this talk, I will give a summary of our works on big media data search and management. In particular, I will present our recent works on compact coding for large scale similarity search: composite quantization for approximate nearest neighbor search (ICML 2014) and sparse composite quantization (CVPR 2015).

Biography

Jingdong Wang is a Lead Researcher at the Visual Computing Group, Microsoft Research Asia. He received the M.Eng. and B.Eng. degrees in Automation from the Department of Automation, Tsinghua University, Beijing, China, in 2001 and 2004, respectively, and the PhD degree in Computer Science from the Department of Computer Science and Engineering, the Hong Kong University of Science and Technology, Hong Kong, in 2007. His areas of interest include computer vision, machine learning, pattern recognition, and multimedia computing. At present, he is mainly working on the Big Media project, including large-scale indexing and clustering, Web image search and mining. He has published 100+ papers, including one single-authored book, book chapters, and other papers in top conferences and prestigious international journals such as CVPR, ICCV, ACMMM, ICML, SIGIR, TPAMI, IJCV, TIP, TOG (Siggraph), and so on.
He has been served as an area chair in ACMMM 2015, a track chair in ICME 2012, a special session chair in ICMR 2014, an area chair in ICME 2014, a program committee member or a reviewer in top conferences and journals, including CVPR, ICCV, NIPS, SIGIR, SIGGRAPH, and ACMMM, TPAMI, IJCV, TIP, TKDE, TMM, ToMM. He has also been invited to serve as an editorial board member in the international journal of multimedia tools and applications, an associate editor of the international journal of Neurocomputing.

Congratulations to the PALM graduates Yin Chao who won the Jiangsu Province outstanding master's degree thesis!

祝贺PALM毕业生尹超获得江苏省优秀硕士学位论文！

PALM faculties, Prof. X. Geng and Prof. M.-L. Zhang, visited the the School of Computer Science and Engineering, South China University of Technology, on December 14, 2014. Two talks titled "Label Distribution Learning" and "Disambiguation-free Partial Label Learning" are given by Prof. Geng and Prof. Zhang respectively.

2014年12月14日，PALM实验室耿新老师与张敏灵老师应邀访问华南理工大学计算机科学与工程学院，分别做了题为“标记分布学习”以及“非消歧偏标记学习”的学术报告，并与实验室师生进行学术交流。

Three PALM faculties (Prof. X. Geng, Prof. M.-L. Zhang, A/Prof. H. Xue) and two PALM students (Y.-K. Li, D.-Y. Zhang) attended the 13th Pacific Rim International Conference on Artificial Intelligence (PRICAI'14), held at Gold Coast, Australia, December 1-5, 2014. Two oral presentations titled "Enhancing Binary Relevance for Multi-Label Learning with Controlled Label Correlations Exploitation" and "Lp-Norm Multiple Kernel Learning with Diversity of Classes" are given by Y.-K. Li and D.-Y. Zhang respectively.

2014年12月1-5日，PALM实验室耿新、张敏灵、薛晖、李宇琨、张大银参加于澳大利亚黄金海岸召开的第13届亚太人工智能大会，李宇琨与张大银分别做了题为 “Enhancing Binary Relevance for Multi-Label Learning with Controlled Label Correlations Exploitation” 以及 “Lp-Norm Multiple Kernel Learning with Diversity of Classes” 的口头报告。

PALM faculties, Prof. X. Geng and Prof. M.-L. Zhang, visited the MOE Key Laboratory of Computational Intelligence and Chinese Information Processing, Shanxi University, during November 14-15, 2014. Two talks titled "Label Distribution Learning" and "Learning with Weak Supervision" are given by Prof. Geng and Prof. Zhang respectively.

2014年11月14-15日，PALM实验室耿新老师与张敏灵老师应邀访问山西大学计算智能与中文信息处理教育部重点实验室，分别做了题为“标记分布学习”以及“弱监督学习”的学术报告，并与实验室师生进行学术交流。

2014年11月7日，第十二届中国机器学习及其应用研讨会（MLA）在西安电子科技大学举行。来自全国各大高校、研究所的众多机器学习领域的专家学者参加了此次会议。MLA是由陆汝钤院士发起组织的“智能信息处理系列研讨会”，2002年12月在复旦大学智能信息处理上海市重点实验室举办了第一届。

PALM实验室耿新老师、张敏灵老师、刘胥影老师以及七位同学参加了此次会议。张敏灵老师担任了此次大会部分主持工作，耿新老师应邀做了精彩报告，讲授了自己对于CVPR的一些见解以及论文写作、投稿需要注意的地方。

李宇琨同学的海报《Enhancing Binary Relevance for Multi-Label Learning with Controlled Label Correlations Exploitation》以及罗龙润同学的《Multilabel Ranking with Inconsistent Rankers》参加了墙展交流。

张老师为大会主持人。

罗龙润及其post。

罗龙润为会议参与人介绍他的研究内容。

耿老师介绍CVPR会议。

于菲及post。

于菲为会议参与人介绍李宇琨的研究内容。