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

A Freudenthal triple system (FTS) is a vector space endowed with a quartic form and a bilinear form such that a triple product defined from these forms satisfies a specific identity. The original example is the 56-dimensional representation of E_7; here, the group stabilizing both forms is precisely E_7. M. Rost observed that an 8-dimensional vector space with quartic form occurring in a paper of M. Bhargava was, with a suitable bilinear form, a FTS; he asked what the stabilizer of the forms was in this case. We answer his question by showing that both his example and the 56-dimensional representation of E_7 are instances of a general construction that reveals a FTS within any Lie algebra of type B, D, E or F, with natural definitions for the quartic and bilinear forms.

One day, a little French boy named Henri was playing in a field with his boomerang. He decided to try an experiment: ``what would happen," he thought, ``if I tied a very long string to my boomerang and then held on to one end. When the boomerang comes back," he wondered, ``will I be able to pull the string back to me while holding on to both ends?" ``Surely," he thought, ``as long as it doesn't get `caught' on something I can pull it back. But can it get caught on space itself?" Little Henri Poincaré probably never owned a boomerang, but one can imagine he may have had similar thoughts: his conjecture that any space where such a mathematical string never gets caught must be the three-dimensional sphere mystified mathematicians for nearly a hundred years. That is until a reclusive Russian named Grisha Perelman quietly posted three preprints on a public web server that settled Poincaré's question once and for all. Perelman's proof, a tour-de-force of three-dimensional geometry, presents as many new challenges as solutions. In this talk, I'll attempt to untangle elements of the history and conclusion of this elusive conjecture, and where we go from here. Along the way, we'll learn of non-Euclidean `Escher-esque' tilings, triangles whose interior angles don't add up to 180 degrees, how our universe is really just a pair of multi-handled coffee cups. The talk will be aimed at undergraduates but should be understandable to a lay audience.

The study of objects defined as fixed points under a Galois action is a big part of the algebra program at Emory. We develop the basic setup, using nothing but 1st year graduate algebra.

Suppose you are given four cards, each containing four nonnegative real numbers, written one below the other, so that the sum of the numbers on each card is 1. You are allowed to put the cards in any order you like, then you write down the first number from the first card, the second number from the second card, the third number from the third card, and the fourth number from the fourth card, and you add these four numbers together. What is the smallest interval [a, b] so that, no matter which cards you are given, there is always an ordering of the cards so that the sum will lie in [a, b]? I will give the history of this problem and what I know about it.

The World Wide Web is rapidly evolving into a "social Web" that is dominated by user-generated content and user-centric social interactions. Although this evolution enhances the usability of the Web, it also presents a number of information security and privacy challenges. Some of these challenges are familiar such as malware and spam, while others are new and unique to the social Web such as fraudulent user profiles. In this talk, we survey the landscape of threats facing users in the social Web, and we explore interesting research challenges that arise when countering these threats. The threats facing social Web users can be broadly categorized into three distinct classes: traditional attacks, socially enhanced attacks, and social Web-specific attacks. In the talk, we will discuss each of these attack classes and examine numerous examples that have been observed over the past few years. These examples are real-world incidents that have already affected specific social Web environments (e.g., MySpace, Facebook, etc.), highlighting the practical importance of these attacks and their impact on millions of social Web users. We will also investigate various countermeasures for these threats and discuss important research questions that remain unanswered.\\ \\ Bio:\\ \\ Dr. Steve Webb is a research scientist at Purewire, Inc. and the chief architect of PurewireTrust.org, a free online portal that maintains portable reputations for people, places, and things. Dr. Webb has 10 years of experience in researching security threats that target information systems such as P2P networks, email systems, the World Wide Web, and social networking environments. He holds a Ph.D. in Computer Science from the Georgia Institute of Technology, where he published more than a dozen academic articles on various information security and privacy issues, including spam, phishing, and information warfare.

Role based access control is well accepted as the standard best practice for access control within applications and organizations. Role engineering, the task of defining roles and associating permissions to them, is essential to realize the full benefits of the role-based access control paradigm. The essential question is how to devise a complete and correct set of roles -- this depends on how you define goodness/interestingness (when is a role good/interesting?) We define the role mining problem (RMP) as the problem of discovering an optimal set of roles from existing user permissions. In addition to the above basic RMP, we introduce two different variations of the RMP, called the delta-approx RMP and the Minimal Noise RMP that have pragmatic implications. Our main contribution is to formally define RMP, analyze its theoretical bounds, and present heuristic solutions to find the optimal set of roles based on subset enumeration. We place this in the framework of matrix decomposition which is applicable to many other domains including text mining. \\ \\ Bio:\\ \\ Dr. Vijay Atluri received her Ph.D. in Information Technology from George Mason University, USA. She is currently a Professor of Computer Information Systems in the MSIS Department, and research director for the Center for Information Management, Integration and Connectivity (CIMIC) at Rutgers University. Dr. Atluri's research interests include Information Systems Security, Privacy, Databases, Workflow Management, Spatial Databases, Multimedia and Distributed Systems. Currently, she serves as a member of the Steering Committee and the secretary/treasurer for ACM Special Interest Group on Security Audit and Control (SIGSAC), and chair of the IFIP WG11.3 on Data and Application Security.

Let $C$ be a curve over a $p-$adic field $F$ and $K = F(C)$. For division algebras of exponent prime to $p$, it is known that index divides the square of the exponent and division algebras of prime degree are cyclic. Both results avoid the prime $p$ because in that case there is no good theory of ramification of Brauer group elements. However, one can try and avoid this obstacle by defining the ramification group of a discrete valued field $K$ with valuation ring $R$ as the quotient of Brauer groups $Br(K)/Br(R)$ and then study the functorial properties of this quotient. One is then led to the complete case and to consider the paper ``A generalization of local class field theory by using K groups I'' by Kazuya Kato (J Fac Sci Sec. IA 26, 2 303-376). We will discuss the progress we have made on this problem using Kato's work.