大会报告（Keynotes）
数据缓存管理技术的研发进展
张晓东 美国俄亥俄州立大学
张晓东是美国俄亥俄州立大学的 Robert M. Critchfield讲席教授，并担任计算机科学与工程系主任。他主持研究的一些核心算法和系统设计已被广泛应用到商业处理器和主要操作系统和数据库系统中，有效地优化或更新了计算机存储系统中的一些关键技术。张晓东在北京工业大学获电气工程学士学位,在美国科罗拉多大学获计算机科学博士学位。他是国际电气电子工程师协会 (IEEE) Fellow，并获得2010年中国计算机学会颁发的海外杰出贡献奖。
报告摘要：
Cache replacement algorithms play a critical role in buffer management in all memory-capable digital systems from large computer and database systems to small cell phones. After our concentrated efforts for several years, the dominant LRU-like algorithms have been gradually replaced by LIRS-like algorithms in major operating and data database systems. In this talk, I will present how LIRS algorithm and its variations fundamentally address the limits of LRU, and how efficient system implementations turn our research results into reality in production systems. Four connected technical components in the presentation include: (1) the LIRS algorithm, (2) approximations and variations of the algorithm in practice, (3) lock-contention free implementations, and (4) software cache partitioning mechanism in multi-core processors.
How to Share Data Securely
Dr. X. Sean Wang 王晓阳 The University of Vermont and US National Science Foundation
Dr. X. Sean Wang is currently a Program Director at the National Science Foundation (NSF) in the Division of Intelligent Information Systems. At NSF, he helps manage the Information Integration and Informatics program and the Trustworthy Computing program. He's on a detail assignment to the NSF from the University of Vermont, where he is the Dorothean Professor in Computer Science in the College of Engineering and Mathematical Sciences. He earned his PhD degree in Computer Science from the University of Southern California in 1992. Dr. Wang's research areas include Database Management Systems and Cyber Security and Privacy.
报告摘要：
Data are precious assets to many people and organizations. Protection of data has been an important topic ever since we started to collect them. But data protection in the 21st century has become more complicated due to the ever increasing needs and benefits of sharing data. An interesting observation is that data semantics play an important role. In this talk, I will give a personal view on the myriad research aspects that need our attention, and discuss various approaches.
孔子与他聪明的弟子
张智威 （Edward Chang） 谷歌中国研究院副院长
张智威教授于2006年加入Google（谷歌）公司，担任中国研究院副院长，负责技术研究与社区类产品团队组建与研发工作。2010年起并担任谷歌全 球移动研发计划Mobile 2014 负责人， 主持新一代移动技术创新研发。
      1999年8月，获得斯坦福大学电机工程博士学位之后，张智威教授受聘于加利福尼亚大学圣巴巴 拉分校电子和计算机工程系。2003年3月，他获得终身职位，并于2006年晋升为电子工程学正教授。张智威教授是机器学习和多媒体研究领域的世界级专家。他最近的研究集中在大规模机器学习、数据挖掘、及高维度数据检索。他领导的小组在以下领域卓有建树：借助内核方法，通过积极学习掌握图像、视频搜索概念；通过动态联想和内核定向规划远程功能；视频传感器数据的管理和融合；高维度图像/视频信息的分类和检索；以及并行矩阵因数分解 和加速支持向量机的运转。
      张智威教授在国际计算机协会（ACM）、美国电气及电子工程师学会（IEEE）和美国工业 与数学应用协会（SIAM）等多个组织的会议程序委员会任职。他共同发起了ACM视频传感器网络研讨年会，并从2003年起担任年会的联合主席。近年，他 共同主持了五项国际会议：多媒体数据建模会议（北京）、影像科学和技术学会（IS＆T）、ACM多媒体会议（圣巴巴拉）、2008 年IEEE和 WWW 学术会议。他是IBM教授奖和美国国家科学基金委颁发的杰出青年奖得主。
报告摘要：
孔子是中国古代伟大的导师。他的理论和原则，有效地由他的弟子传遍中国。”孔子”是谷歌的知识搜索产品代号。这个主题演讲将陈述知识搜索产品的关键”弟子”：几个机器学习的关键技术。这些关键技术产生问题标签，搜索相关问题，评价答案素质，排名用户，提取高质量的答案，配置相关广告等等。这个主题演讲除了陈述算法，也阐述如何使用云计算技术处理庞大数据集。
模糊对象的最近邻查询（K-Nearest Neighbor Search for Fuzzy Objects）
周晓方 澳大利亚昆士兰大学
周晓方，澳大利亚昆士兰大学计算机科学教授，中国人民大学“千人计划”特聘教授，昆大数据与知识工程研究室主任，澳洲研究院(ARC)企业信息基础研究网络主任。周晓方教授长期从事数据库系统和信息系统研究，在大规模复杂数据管理，集成与分析的基础研究及海量数据前端应用方面积累了大量经验和成果，其主要研究领域包括空间数据库，多媒体数据库，数据质量，网络信息系统及高性能数据处理。周晓方教授是VLDB Journal, IEEE Transactions on Knowledge and Data Engineering和World Wide Web Journal等国际一流学术期刊的编委。
报告摘要：
The K-Nearest Neighbor search (kNN) problem has a broad range of applications. In this talk we will present a new version of this problem in the context that both the target data and the query object are fuzzy (i.e. with indeterministic boundaries). Fuzzy objects play an important role in many application areas such as biomedical imaging and GIS. Existing work on fuzzy objects mainly focus on defining basic fuzzy types and operations, leaving query processing for advanced queries such as kNN largely untouched. In this talk, we will propose a distance function that can be used to measure the distance between two fuzzy objects, and identify two new types of kNN queries for fuzzy objects, Ad-hoc kNN query (AKNN) and Range kNN query (RKNN), which find the k nearest objects qualifying at some probability threshold or within a range of probability thresholds respectively. For AKNN query processing, we derive a tight lower bound and upper bound for a distance function between fuzzy objects and extend the R-tree index to reduce the search space. To improve the performance of RKNN search, a number of theoretical results are developed to filter out a large amount of disqualifying objects as early as possible.