学术讲座通知​:城市计算与大数据

讲座题目:城市计算与大数据

主讲人:郑宇博士(微软亚洲研究院主管研究员)

主持人:马占宇副教授

时间:2015年11月11日(周三)14:30~16:00

地点:教三楼811学术报告厅

摘要:城市计算是计算机科学以城市为背景,跟城市规划、交通、能源、环境、社会学和经济等学科融合的新兴领域。城市计算通过不断获取、整合和分析城市中不同领域的大数据来解决城市所面临的挑战。本报告将概述城市计算的定义、框架和主要研究问题,以典型应用为案例着重介绍大数据中跨域数据的融合和协同计算技术。具体案例包括基于大数据的细粒度空气质量分析和预测、城市油耗和汽车尾气排放评估,以及城市住房价值分级和评估等。相关技术发表在KDD等顶尖国际会议和期刊上,并在实际系统中部署应用。更多细节请看http://research.microsoft.com/en-us/projects/urbancomputing/default.aspx。

 

 

 

校学术委员会

信通院

2015年11月4日

学术报告: Complete Dictionary Recovery over the Sphere

报告题目: Complete Dictionary Recovery over the Sphere
报告人:Ju Sun, Electrical Engineering, Columbia University
主持人: 北邮模式识别实验室 李春光
时间:2015年9月2日 15:00-16:30
地点:教三810会议室

报告摘要:We consider the problem of recovering a complete (i.e., square and invertible) matrix $A_0$, from $Y \in R^{n\times p} with $Y = A_0X_0$, provided $X_0$ is sufficiently sparse. This recovery problem is central to the theoretical understanding of dictionary learning, which seeks a sparse representation for a collection of input signals, and finds numerous applications in modern signal processing and machine learning. We give the first efficient algorithm that provably recovers $A_0$ when $X_0$ has $O (n)$ nonzeros per column, under suitable probability model for $X_0$. In contrast, prior results based on efficient algorithms provide recovery guarantees when $X_0$ has only $O (pn)$ nonzeros per column. Our algorithmic pipeline centers around solving a certain nonconvex optimization problem with a spherical constraint, and hence is naturally phrased in the language of manifold optimization. To show this apparently hard problem is tractable, we first provide a geometric characterization of the high-dimensional objective landscape, which shows that with high probability there are no “spurious” local minima. This particular geometric structure allows us to design a Riemannian trust region algorithm over the sphere that provably converges to one local minimizer with an arbitrary initialization, despite the presence of saddle points. The geometric approach we develop here may also shed light on other problems arising from nonconvex recovery of structured signals.

报告人简介:Ju Sun is now a five year PhD candidate (Advisor: Prof. John Wright) in the Department of Electrical Engineering, Columbia University in the City of New York. He works at the intersection of computer vision, machine learning, numerical optimization, signal/image processing, information theory, and compressive sensing, focusing on modeling, harnessing, and computing with structures in massive data, with provable guarantees and practical algorithms. His paper “Complete Dictionary Recovery over the Sphere” has recently received the best student paper award from SPARS 2015, which was held in Cambridge University.

详细信息请参阅
http://www.columbia.edu/~js4038/
http://sunju.org/

欢迎各位老师和同学积极参加!

图像识别技术其智能应用

主讲人:蒋树强,中科院计算所研究员,博士生导师

时间:201571日(周三)下午14:00-16:00

地点:教三楼811会议室

主持人:马占宇  副教授 北邮模式识别实验室

内容摘要:

自动图像识别是计算机视觉和多媒体领域的重要研究目标之一,具有广阔的应用前景。由于受到一义多图、一图多物、一物多态、异物相似等多重因素的影响,图像识别技术仍存在许多悬而未决的问题,距实际应用仍有很大差距;如何面向具体应用建立图像的高效表示和分类方法、实现图像的准确识别是一项值得研究的课题。本报告将从纯视觉信息的图像识别和基于上下文信息的图像识别两个角度进行技术介绍,包括实例级物体识别、场景识别、基于地理上下文的图像识别和基于RGB-D的手持物体识别,这些技术可以分别应用在手机图像识别和面向机器人的视觉交互等。报告最后对应用前景进行展望并进行技术演示。

主讲人简介:

蒋树强,中科院计算所研究员,博士生导师,IEEE Senior Member,研究方向为图像/视频等多媒体信息的分析、理解与检索技术,共在国内外刊物和会议上发表论文100多篇,获授权专利10项,部分技术被应用于多个实际系统中,获得2008年度北京市科技新星计划支持,2012年度中国科学院卢嘉锡青年人才奖,2012年度中国计算机学会科学技术奖(排名第二),2013年度中国科学院青年科学家国际合作奖,获2013年度国家自然科学基金优秀青年科学基金支持。

学术讲座通知:Deep models for face processing with “big” or “small” data

主讲人:山世光,中国科学院计算技术研究所研究员,中科院智能信息处理重点实验室常务副主任

时间:2015618日(周四)上午9:00-11:00

地点:教三楼811会议室

主持人:马占宇  副教授 北邮模式识别实验室

内容摘要:

Deep learning models, especially CNN, has been successfully applied to face recognition, especially under the evaluation protocol of Labeled Faces in the Wild (LFW), when big face data is available. In this talk, except showing some recent results of CNN feature for video-based face processing (our FG’15 paper), I will also show that alternative deep models such as Auto-Encoder can also benefit face recognition impressively, especially for face alignment (our ECCV14 paper) and pose normalization (our CVPR14 paper) purpose. Both works might imply in case of “small” data, elaborate deep models can also work well for many computer vision tasks.

主讲人简介:

山世光,中国科学院计算技术研究所研究员、博士生导师,中科院智能信息处理重点实验室常务副主任。主要从事图像处理与理解、计算机视觉、模式识别、智能人机交互界面等相关研究工作。已在国际/国内期刊、国际会议上发表/录用学术论文200余篇,其中CCF A类国际会议和期刊论文40余篇。论文曾获CCF A类国际会议CVPR2008大会颁发的Best Student Poster Award Runner-up奖。所发表论文被国内外同行引用6800余次(Google Scholar),领导课题组完成的人脸识别系统多次获得国内外人脸识别竞赛第一名。应邀担任CCF-A类国际刊物IEEE Trans. on Image Processing以及Neurocomputing ,EURASIP Journal of Image and Video Processing, Frontier of Computer Science, 《计算机研究与发展》等期刊的编委(Associate Editor),应邀担任过ICCV2011, ICPR2012, ACCV2012, FG2013, ICASSP2014和ICPR2014等相关领域重要国际会议的Area Chair(领域主席)。所完成的人脸识别研究成果2005年度国家科技进步二等奖(第3完成人)。他是2012年度国家自然科学基金委员会首届“优青”获得者。

该讲座为北京邮电大学60周年校庆系列讲座之一。

学术讲座通知

香港中文大学的Tan Lee副教授和Ken Ma副教授应高等智能与网络服务111引智基地的邀请,将于12月15日来北邮作学术报告。欢迎有兴趣的师生踊跃参与。

报告一:

题目:Unsupervised Acoustic Modeling for Spoken Language Applications

主讲人:Tan Lee 副教授

主持人:王小捷 教授

时间:20141215日(星期一)上午9:00-10:00

地点:教三楼811会议室

报告摘要:

Acoustic modeling is an important problem in many spoken language applications. It aims at providing compact yet accurate statistical representations for a set of sub-word units. Conventional acoustic modeling is a highly supervised process that requires plenty of speech data with transcriptions. Such resources may not be available for many languages and in many real-world situations. In this lecture, a new framework of unsupervised acoustic modeling is presented. Different types of posterior features are proposed as segment representations. Spectral clustering algorithms are applied to group short speech segments into phone-like units. The resulted acoustic models can be used in many spoken language applications, including spoken term detection, language recognition, and topic identification.

主讲人简介:

Tan Lee is an Associate Professor at the Department of Electronic Engineering, the Chinese University of Hong Kong (CUHK). He has been working on speech and language related research since early 90s. His works cover many different areas, including automatic speech and speaker recognition, text-to-speech, speech enhancement, language identification, pathological speech analysis, hearing and speaking aids, and music signal processing. Tan Lee initiated and coordinated a number of pioneering projects on the research and development of Chinese spoken language technologies in Hong Kong. He led 8 projects funded by the General Research Funds (GRF) from the Hong Kong Research Grants Council (RGC). Tan Lee works closely with medical doctors, and speech and hearing professionals, in applying signal processing techniques to human communication disorder problems. He is the Director of the newly established Language and Communication Disorders Research Laboratory at CUHK Shenzhen Research Institute. Tan Lee was the Chairman of the IEEE Hong Kong Chapter of Signal Processing in 2005-2006. He is an associate editor of the EURASIP Journal on Advances in Signal Processing. Tan Lee received the CUHK Vice-Chancellor’s Exemplary Teaching Award in 2004.

报告二:

题目:Hyperspectral Unmixing in Remote Sensing: What Do Signal Processing People Learn from There?

主讲人:Ken Ma副教授

主持人:马占宇

时间:20141215日(星期一)上午10:00-11:00

地点:教三楼811会议室

报告摘要:

The aim of this talk is to overview hyperspectral unmixing (HU) techniques from a signal processing researcher’s perspective. HU is one of the most prominent research topics in hyperspectral remote sensing. The problem is to identify materials and their corresponding compositions in a captured scene, using the high spectral degrees of freedom of hyperspectral sensors. From a signal processing viewpoint, this is a blind source separation (BSS) problem. We will review how clever insights from remote sensing researchers and recent involvements from other fields, such as signal processing, optimization and machine learning, lead to elegant HU theory and methods – which depart quite significantly from conventional BSS techniques, and in fact, give new insights to BSS theory and methods. The connections of HU to other areas, such as text mining, biomedical imaging and computer vision, may also be discussed, depending on the availability of time.

主讲人简介:

Wing-Kin (Ken) Ma is currently an Associate Professor with the Department of Electronic Engineering, The Chinese University of Hong Kong. His research interests are in signal processing and communications, with recent activities focused on optimization, MIMO transceiver designs and interference management, blind signal processing theory, methods and applications.

Dr. Ma is active in the Signal Processing Society. He is currently serving or has served Associate Editor and Guest Editor of several journals, which include IEEE Transactions on Signal Processing, Signal Processing, IEEE Journal of Selected Areas in Communications and IEEE Signal Processing Magazine. He is a Member of the Signal Processing Theory and Methods (SPTM) Technical Committee. His students won ICASSP Best Student Paper Awards in 2011 and 2014, respectively. He was a tutorial speaker of EUSIPCO 2011 and ICASSP 2014, respectively.

学术讲座通知

瑞典计算机科学学院(SICS-Swedish Institute of Computer Science) 的Anders Lindgren和Fehmi Ben Abdesslem博士应高等智能与网络服务111引智基地的邀请,将于10月22日来北邮作学术报告。欢迎有兴趣的师生踊跃参与。

报告一:

题目:Keeping it real – The importance of experiencing and understanding real mobile network environments

主讲人:Anders Lindgren研究员

主持人:马占宇

时间:20141022日(星期三)上午10:00-11:00

地点:教三楼811会议室

主讲人简介:

Dr. Anders Lindgren is a senior researcher at SICS Swedish ICT and an adjunct lecturer at Luleå University of Technology. He received his Ph.D. from Luleå University of Technology in 2006, and worked as a postdoctoral researcher at University College London and University of Cambridge. His research interests include opportunistic and information-centric networking, communication and computation in challenged environments, efficient IoT, and big data analytics for mobile networks.

报告摘要:

Over the past decade, mobile networking has seen an explosive growth, in particular with the introduction of smart-phones. This is creating incredible new opportunities for users to stay connected and utilize our connected society in their daily life. It does however also come with problems as operator networks become overloaded, without operators necessarily understanding exactly how their customers will use the network. Furthermore, it also comes with the risk of further increasing the digital divide between the rich and poor parts of the world, as some societal services will only be available to those with access to connectivity.

In this talk, I will focus on the importance of keeping it real, and not solely design networks based on synthetic models and simulations. Through examples from real deployments, field experiences, and network measurements, I will show some examples of what users really want in different scenarios and this can be realized even in challenged scenarios.

Mobile networks span a very wide range of scenarios and use cases, and I will focus at some particular scenarios at the extreme ends of this spectrum. I will explain how they differ and are similar to more common situations, and discuss why it is important to understand the specific characteristics of these situations and how they affect network design. In particular, I will look at the following two scenarios based on previous personal experience:

– What can be done for users that do not have access to reliable (or low-capacity) communication infrastructure? I will explain some existing technical solutions for providing network access to people living outside the rich and well-connected parts of the world, and will, most importantly, explain the importance of properly understanding the target user groups in order to create good network designs.

报告二:

题目:Big Data in Computer Networks – From Mobile Devices to The Internet of Things

主讲人:Fehmi Ben Abdesslem研究员

主持人:乔媛媛

时间:20141022日(星期三)上午11:00-12:00

地点:教三楼811会议室

主讲人简介:

Dr Fehmi Ben Abdesslem received his M.Sc and PhD from the University of Paris 6 in 2008, before working as a research associate at the University of St Andrews, and at the University of Cambridge. He has then been awarded a Marie-Curie research fellowship from the European Commission (ERCIM) to join SICS, and is now a permanent Senior Research Scientist at the Decisions Networks and Analytics laboratory.

报告摘要:

Big data is nowadays a common term regularly appearing in the news as part of our every day life. Large amounts of data are constantly collected, stored and analysed. In this presentation, we will first explain and discuss this trend before highlighting the research challenges in computers networks. We will then describe an example of big dataset collected from mobile phones, and show the applications of analysing such datasets. Finally, we will provide a short introduction about the potential benefits of the big data generated by the Internet of things.

学术讲座:多媒体数据内容分析与保密防范

主讲人:操晓春研究员,中国科学院信息工程研究所信息安全国家重点实验室

时间:20141016日(周四)上午9:00-11:00

地点:教三楼811会议室

主持人:马占宇

内容摘要:

报告人将汇报其团队在图像视频的事件检测与识别、物体分割、编辑篡改、主动保护、篡改恢复、取证、信息隐藏、隐写分析、防光学采集、防窃拍窃录等方面的研究进展。
主讲人简介:

操晓春:中国科学院信息工程研究所信息安全国家重点实验室研究员、博导,2012年入选中科院“百人计划”。本科和硕士就读于北京航空航天大学,博士毕业于美国中佛罗里达大学。曾就职于美国ObjectVideo公司和天津大学。主要从事信息安全和计算机视觉领域的研究,取得了多项创新研究和实践成果,并获得重要应用。曾先后担(兼)任中国计算机学会青年工作委员会秘书、中国计算机学会青年计算机科技论坛(YOCSEF)学术委员会副主席、中国计算机学会会员与分部工作委员会常务委员。操晓春研究员还获得了包括国家自然基金重点项目和优秀青年基金项目在内的多项项目资助。

学术讲座: Event Reasoning and Game theory for Decision Support under Uncertainty: with applications in surveillance

主讲人:Weiru Liu教授,Queen’s University Belfast

时间:2014911日(周四)上午9:00-10:00

地点:教三楼811会议室

主持人:马占宇北邮模式识别实验室

内容摘要:

Three issues are usually associated with large sensor network based intelligent surveillance systems. First, the collection, analysis, fusion and interpretation of a large amount of incomplete and heterogeneous information. Second, the demand of effectively predicting suspects’ intentions and ranking the potential threats posed by each suspect. Third, strategies of allocating limited security resources (e.g., the dispatch of security team) to prevent a suspect’s further actions towards critical assets. In this talk, I will present a multi-agent based event reasoning framework that can detect, correlate and reason with elementary events to predict potential threats. Game theory is then applied to select best strategies to allocate security resources based on threat degrees and suspects’ intentions.

Prof. Liu will also take this opportunity to give an overview of major research topics and future research directions in the Knowledge and Data Engineering Cluster, within the School of Electronics, Electrical Engineering and Computer Science, Queen’s University Belfast.

 

 

主讲人简介:

Prof. Weiru Liu holds the Chair of Artificial Intelligence, and is the Director of Research of the Knowledge and Data Engineering Cluster in the School of EEECS at QUB. Her research interests include noisy and uncertainty information modeling and fusion for sensor networks with applications to cyber and physical security, smart homes; intelligent autonomous systems, especially on agent’s belief revision with uncertain observations; game-theoretic resource management in multi-agent systems; and large-scale data analytics with particular focus on anomaly detection. She has over 140 publications, was the Conference and Program Chair of the 11th European Conference on Symbolic and Quantitative Reasoning under Uncertainty (ECSQARU’2011), was Program Co- Chair of 7th Int. Conf. on Scalable Uncertainty Management (SUM’2013). Her research has been supported by UK EPSRC, UK ESRC, UK TSB, the Royal Society, EU, and industry with total funding value over £17Million (projects of which she was/is either Principal Investigator or Co-Investigator) and had won EPSRC QUB Research Impact Award in 2011.

学术讲座: Learning representations for semantic relational data

学术讲座通知

讲座题目: Learning representations for semantic relational data

主讲人:Patrick Gallinari教授

时间:2014年7月11日(周五)下午15:00-17:00

地点:教三楼811会议室

主持人:高升 模式识别实验室

内容摘要:Learning representations for complex relational data has emerged at the crossroad between different research topics in machine learning. The motivation of this work is often driven by the applications themselves and by the nature of the data which are often complex (multimodal, heterogeneous, dynamic), and multi-relational (e.g. biology, social networks). One possible approach is to map these data onto one or more continuous latent spaces in order to obtain representations on which it is possible to use classical machine learning methods. In recent years, several lines of research have developed these ideas, sometimes independently, and they are now represented in the “Learning Representations” community. The tools deployed rely on statistical modeling, on linear algebra with matrix or tensor factorization, or more recently on neural networks. The presentation will give a brief presentation of some of these methods and show applications in the field of semantic data analysis and social networks.

主讲人简介:Patrick. Gallinari is professor in Computer Science at Universite Pierre et Marie Curie (UPMC), France. His research domain is primarily statistical machine learning with applications to domains involving semantic data like information retrieval. His recent work has focused on statistical modeling of complex relational data described by sequences, trees or graphs.  Before that, he has been a pioneer of neural networks in France, participating to the development of this domain in Europe. He has also been director of the computer science lab. at UPMC for about 10 years.

 

学术报告: Subspace Clustering – Recent Advances

报告题目:Subspace Clustering – Recent Advances
报告人:Zhouchen Lin(林宙辰)博士 北京大学教授
时间:2014年4月10日(周四)09:30-11:00
地点:教三 810会议室

报告摘要: Nowadays we are in the big data era, where the data is usually high dimensional. How to process high dimensional data effectively is a critical issue. Fortunately, we observe that data usually distribute near low dimensional manifolds. Mixture of subspaces is a simple yet effective model to represent high dimensional data, where the membership of the data points to the subspaces might be unknown. Therefore, there is a need to simultaneously cluster the data into multiple subspaces and find a low-dimensional subspace fitting each group of data points. This problem, known as subspace clustering, has found numerous applications. In this talk, I will present my recent work on this research problem.

报告人简介:
ZHOUCHEN LIN received the Ph.D. degree in applied mathematics from Peking University in 2000. He is currently a Professor at Key Laboratory of Machine Perception (MOE), School of Electronics Engineering and Computer Science, Peking University. He is also a Chair Professor at Northeast Normal University and a guest professor at Beijing Jiaotong University. Before March 2012, he was a Lead Researcher at Visual Computing Group, Microsoft Research Asia. He was a guest professor at Shanghai Jiaotong University and Southeast University, and a guest researcher at Institute of Computing Technology, Chinese Academy of Sciences. His research interests include computer vision, image processing, computer graphics, machine learning, pattern recognition, and numerical computation and optimization. He is an associate editor of International J. Computer Vision and a Senior member of the IEEE. He served CVPR2014 as an area chair.