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

Missing web pages (pages that return the "404 Page Not Found" error) are part of the browsing experience. So too are pages whose owners failed to renew their domain and whose old URls now have unexpected content. Users that encounter a missing page or unexpected page may try to use search engines to discover either the same page at a new location or a similar, "good enough" page to satisfy their information needs, but this can be laborious. To address this need, we are developing a semi-automated framework to assist users to first discover the topic of the missing page, and then locate the same or similar page at a new URl.\\ \\ We have been investigating a number of techniques to discover the "aboutness" of an unknown web page. If the page is in the internet Archive's Wayback Machine or in a search engine cache, the user may be satisfied with the old copy. If an old copy is insufficient, we can use either the page's title or generate a lexical signature to serve as a queiry to a search engine to find the resource. A lexcial signature is a 5-7 word "abstract" of a document that is suitable for using as a queiry to a search engine. The performance of titles and lexcial signatures are comparable, with both achieving over 60 percent success. The combination of titles and lexical signatures from link neighborhoods as well as using tags from del.icio.us, but at this point neither method performs well.\\ \\ Speaker Bio:\\ Michael L. Nelson is an associate professor of computer science at Old Dominion University. Prior to joining ODU, he worked at NASA Langley Research Center from 1991-2002. He is co-editor of the OAI-PMH and OAI-ORE specifications and is a 2007 recipient of an NSF CAREER award. In 2008, Dr. Nelson was named a "Digital Preservation Pioneer" by the Library of Congress. He has developed many digital libraries, including the NASA technical Report Server. His research interests include repository-object interaction and alternative approaches to digital preservation. More information about Dr. Nelson can be found at http://www.cs.odu.edu/~min/

In recent years, the framework of Monge-Kantorovich optimal transport has proven quite successful in the design of innovative optical systems. A major improvement from a theoretical point of view as well as for practical applications is the design of free-form lenses and mirrors. The optimal transport approach has been applied to a variety of design problems for systems consisting of one or two lenses or reflectors. A classical problem in optical design is the so-called illumination or near-field single reflector problem. In this problem, one needs to design a reflecting surface such that the rays emitted from a non-isotropic source are reflected in such a way to create a prescribed in advance illumination pattern on a given target set at a finite distance. In this talk, we will discuss how the near-field problem fits into a generalized optimal transport framework and how solutions to this transport problem are related to solutions of the near-field problem.

Recently, M. Marshall answered a long-standing question in real algebraic geometry by showing that if $f(x,y) \in \mathbf{R}[x,y]$ and $f(x,y) \geq 0$ on the strip $[0,1] \times \mathbf{R}$, then $f$ has a representation $f = \sigma_0 + \sigma_1 x(1 - x)$, where $\sigma_0, \sigma_1 \in \mathbf{R}[x,y]$ are sums of squares. In this talk, we give the background to this result, which goes back to Hilbert's 17th problem, and our generalizations to other non-compact basic closed semialgebraic sets of $\mathbf{R}^2$ which are contained in strip. We also give some negative results.

Record Linkage is the task of finding entries that refer to the same entity in two or more data sets. It is useful for joining and cleaning data sets that do not have unique common keys, and is important in many studies in business, public health, sociology and psychology, where reconciling independently collected data sets is often the most important and time-consuming first step. In this talk, we describe the basic concepts of automated record linkage and demonstrate FRIL, an open-source software collaboratively developed at Emory and the CDC that facilitates fast and accurate linkages. FRIL implements an array of standard record matching algorithms as well as a large number of user tunable parameters for improving efficiency and precision. It provides a friendly, easy to use interface and a set of useful visualization tools for quick user feedbacks. Since its release in 2008, FRIL has been widely adopted, including uses by research groups at the National Center on Addiction and Substance Abuse, the Harvard Business School and the NIH. This talk should be of interest to faculty and graduate students whose research involve data collection, integration, and cleansing. Pawel Jurczyk is a Ph.D. Candidate in the Computer Science and Informatics Program at Emory University. His interest includes distributed computing, distributed database query processing, and data/information privacy. He has been an ORISE fellow at the CDC since 2007, and has taught courses in Introduction to Computer Science and Databases.

I'll introduce some basic techniques for estimating determinants of operators. Then I'll show how these techniques are applicable to the problem of counting resonances for perturbations of the hyperbolic plane.

One of the earliest results using the probabilistic method, due independently to Lovasz, Payan and Arnautov, shows that every d-regular, n-vertex graph has a dominating set of size at most (n(1 + log(d+1)))/(d+1). In this talk, we show that if the graph has girth five, one can actually find an independent dominating set of roughly the same size, (n log(d))/d + O(n/d). This extends results of Duckworth and Wormald who studied the problem on random regular graphs, and up to the implied constant is sharp. We further construct examples showing that for irregular graphs the corresponding statement with d replaced by the minimum degree does not hold.