Seminars archive

With the trend of cloud computing, data and computing are moved away from desktop and are instead provided as a service from the cloud. Data-as-a-service enables access to a wealth of data across distributed and heterogeneous data sources in the cloud. It remains a challenge, however, to ensure the privacy, interoperability, and scalability for such services.\\ \\ We designed and developed DObjects, a general-purpose P2P-based query and data operations infrastructure that can be deployed in the cloud and provides access to heterogeneous data sources. The system builds on top of a distributed mediator-wrapper architecture where individual system nodes serve as mediators and/or wrappers and interact with each other in a P2P fashion what guarantees good scalability. As an analogy, the system nodes can be considered as droplets, small elements that provide similar functionality in the cloud. Just as thousands or millions of droplets form a single drop in nature, in cloud computing, groups of droplets that provide similar functionality can form a micro-cloud. Micro-clouds are an integral part of the whole cloud computing system and can provide specific services to users.\\ \\ The dissertation also discusses the novel dynamic query execution engine within the data query infrastructure that dynamically adapts to network and node (or droplet) conditions. The query processing is capable of fully benefiting from all the distributed resources to minimize the query response time and maximize system throughput. In addition to leveraging the traditional distributed query optimization techniques, the (sub)queries are deployed and executed on droplets in a dynamic and iterative manner in order to guarantee the best reaction to network and resource dynamics.\\ \\ Finally, the dissertation presents an extension to the basic DObjects model that enables access to private data that is distributed and needs anonymization. The extension enables droplets to form virtual groups in order to addresses two important privacy issues for the sensitive data: privacy of data subjects and confidentiality of data providers. The dissertation discusses decentralized protocols that enable data sharing for horizontally partitioned databases given these constraints. These protocols can be run by the groups of droplets. Concretely, given a query spanning multiple databases, the query results do not contain individually identifiable information. In addition, institutions do not reveal their databases to each other apart from the query results.

TBA

We outline the general development of a theory of symmetric homology of algebras, an analog of cyclic homology where the cyclic groups are replaced by symmetric groups. This theory is developed using the framework of crossed simplicial groups and the homological algebra of module-valued functors. The symmetric homology of group algebras is related to stable homotopy theory. Two spectral sequences for computing symmetric homology are constructed. The $E_1$ term of one of these relates to a new class of geometric complexes for Coxeter groups. Many open questions remain.

Advances of sensor and RFID technology provide significant new power for humans to sense, understand and manage the world. RFID provides fast data collection with precise identification of objects with unique IDs without line of sight, thus it can be used for identifying, locating, tracking and monitoring physical objects. Despite these benefits, RFID poses many challenges for data processing and management. In this talk, I will discuss our work on temporal and location based RFID data management, and rule-based complex RFID event processing. \\ Fusheng Wang is a Senior Research Scientist at Emory University's Center for Comprehensive Informatics. Dr. Wang received his Ph.D. in Computer Science from University of California, Los Angeles in 2004. Before joining Emory, he was a research scientist and project lead at Siemens Corporate Research. His research areas include heterogeneous scientific data management and integration, high performance biomedical image management systems, temporal data management, RFID data management, XML databases, and collaborative information systems.

Parallel computing has been used for scientific computing applications since the 1960s, when the first supercomputers were developed. However, only recently have these programming paradigms become useful for software running on desktop and notebook computers. In this dissertation we demonstrate the advantage of exploiting modern computer architectures in scientific computing with multithreaded programming in Java for applications in image processing. A significant contribution of this work is an open source, multithreaded high performance scientific computing Java library called Parallel Colt. In addition, on top of Parallel Colt, we have implemented six ImageJ plugins for deconvolution, super resolution, fast Fourier transforms and image cropping. Hence, we are able to provide software to solve important problems in real image processing applications, and which can effectively make use of multi-core CPUs available on affordable desktop and notebook computers.

With the ubiquity of information networks and their broad applications, there have been numerous studies on the construction, and mining of information networks in multiple disciplines, including social network analysis, World-Wide Web, database systems, data mining, machine learning, and networked communication and information systems. However, large graphs may often be disk resident, and such graphs cannot be efficiently processed. In this talk, we examine the issues of on-line analytic processing, summarization and indexing of large graphs. Specifically, the problem of connectivity in the context of massive graphs is considered. In many large communication networks, social networks and other graphs, it is desirable to determine the minimum-cut between any pair of nodes. We will discuss how a connectivity index can be developed for a massive-disk resident graph. A sampling based approach is deployed to create compressed representations of the underlying graph. Trade-off between processing efficiency and accuracy will be shown. Bio: Philip S. Yu is a Professor in the Department of Computer Science at the University of Illinois at Chicago and also holds the Wexler Chair in Information Technology. Before joining UIC, he spent most of his career at IBM Thomas J. Watson Research Center and was manager of the Software Tools and Techniques group. Dr. Yu is a Fellow of the ACM and the IEEE. He served as the Editor-in-Chief of IEEE Transactions on Knowledge and Data Engineering (2001-2004). He is an associate editor of ACM Transactions on Knowledge Discovery from Data and also ACM Transactions of the Internet Technology. He serves on the steering committee of IEEE Int. Conference on Data Mining. He was a member of the IEEE Data Engineering steering committee. Dr. Yu received a Research Contributions Award from IEEE Intl. Conference on Data Mining in 2003. His research interests include data mining, privacy, data stream, and database systems. He has published more than 560 papers in refereed journals and conferences. He holds or has applied for more than 300 US patents. Dr. Yu was an IBM Master Inventor when at IBM.

A Rogers-Ramanujan type identity is a series=product identity which relates certain partitions into parts satisfying difference to partitions into parts satisfying congruence conditions. Rogers-Ramanujan type identities arise in a variety of settings ranging from Lie algebras to statistical mechanics. A supreme example of such a partition identity is the deep 1967 theorem of Gollnitz. We shall discuss a new approach to Gollnitz's theorem in which this partition result and its generalizations will emerge out of a remarkable $q$-hypergeometric identity in three free parameters. This leads to crucial connections with several fundamental results on partitions and $q$-series, a new combinatorial understanding of Jacobi's triple product identity for theta functions, and to some partition congruences modulo powers of 2. The talk will be accessible to non-experts.

Thomason, and Chung, Graham and Wilson were the first to investigate systematically properties of quasirandom graphs. They have stated several quite disparate graph properties -- such as having uniform edge distribution or containing a prescribed number of certain subgraphs -- and proved that these properties are equivalent in a deterministic sense. Simonovits and Sos introduced a hereditary property (which we call S) stating the following: for a small fixed graph L, a graph G on n vertices is said to have the property S if for every subset X of V(G), the number of labeled copies of L in G[X] (the subgraph of G induced by the vertices of (X) is given by $2^{-e(L)} |X|^{v(L)} + o(n^{v(L)})$. They have shown that S is equivalent to the other quasirandom properties. In this talk we give a natural extension of the result of Simonovits and Sos to k-uniform hypergraphs, answering a question of Conlon et al. Our approach yields an alternative, and perhaps simpler, proof of their theorem.

It has recently become known that helicity, defined as the mean streamwise vorticity, is a fundamentally important quantity of the 3D Navier-Stokes equations. Helicity is a conserved quantity of inviscid flow, is cascaded jointly with energy through the inertial range, and is physically interpreted to be the degree to which a flow's vortex lines are tangled and intertwined. Until recently however, helicity has been ignored in NSE computations. In this talk I will show how helicity can be more accurately treated in finite element computations of the 3D NSE, which in turn leads to more accurate simulations. I will discuss two numerical schemes: one which enforces helicity preservation by the NSE nonlinearity (to mimic the continuous case) - joint work with Vince Ervin, and a second that solves for helicity directly inside of a velocity-vorticity method - joint work with Maxim Olshanskii. Numerical examples will be given that demonstrate the effectiveness of the methods.