CSCI 8990 Research Seminar Fall 2009 Abstracts of Talks (in reverse chronological order) -------------------------------------------------------------------------------- Speaker: Charu Vijayagopal Title: Partitioned Scheduling with Fewer Processors When/Where: 12:20--1:10 Tuesday December 8, 2009, room 306 Boyd Abstract: We consider the partitioned Earliest Deadline First (EDF) scheduling of real time periodic or sporadic tasks on identical multiprocessors. Tasks are partitioned using the First Fit Decreasing algorithm, in which tasks are assigned to processors in decreasing order according to their utilization. We characterize our task sets by two parameters: maximum utilization, umax, and utilization binder, gamma. Utilization binder is used to cap the utilization values of tasks. Typically gamma < 1 but in some cases may equal 1. Using gamma we find an upper bound on the value of task utilization that each task in a given task set can have. For a given umax and we have developed a novel method for determining the maximum number of required processors, m(umax, gamma) and utilization bound Ubound(umax, m). Compared to current tests, our method reduces the number of processors by up to 65% and nearly doubles the utilization bound. -------------------------------------------------------------------------------- Speaker: Qi Li Title: SINA -- Scalable Incremental Processing of Continuous Queries in Spatio-temporal Databases When/Where: 12:20--1:10 Friday December 4, 2009, room 306 Boyd Abstract: This paper, authored by Mohamed Mokbel, Xiaopeng Xiong, and Walid Aref, introduces the Scalable INcremental hash-based Algorithm (SINA, for short); a new algorithm for evaluating a set of concurrent continuous spatio-temporal queries. SINA is designed with two goals in mind: (1) Scalability in terms of the number of concurrent continuous spatiotemporal queries, and (2) Incremental evaluation of continuous spatio-temporal queries. SINA achieves scalability by employing a shared execution paradigm where the execution of continuous spatio-temporal queries is abstracted as a spatial join between a set of moving objects and a set of moving queries. Incremental evaluation is achieved by computing only the updates of the previously reported answer. We introduce two types of updates, namely positive and negative updates. Positive or negative updates indicate that a certain object should be added to or removed from the previously reported answer, respectively. SINA manages the computation of positive and negative updates via three phases: the hashing phase, the invalidation phase, and the joining phase. The hashing phase employs an in-memory hash-based join algorithm that results in a set of positive updates. The invalidation phase is triggered every T seconds or when the memory is fully occupied to produce a set of negative updates. Finally, the joining phase is triggered by the end of the invalidation phase to produce a set of both positive and negative updates that result from joining in-memory data with in-disk data. Experimental results show that SINA is scalable and is more efficient than other index-based spatio-temporal algorithms. -------------------------------------------------------------------------------- Speaker: Jeremy Jones Title: Brain-Computer Interfaces When/Where: 12:20--1:10 Friday November 20, 2009, room 306 Boyd Abstract: The talk today will discuss how to automatically customize parameter values for a self-paced brain interface system using an LF-ASD. Previously the parameters were customized by a human expert. The topic paper is titled "Automatic User Customization for Improving the Performance of a Self-Paced Brain Interface System", by Fatourechi, Bashashati, Birch, and Ward. -------------------------------------------------------------------------------- Speaker: Kitae Myoung Title: Question Answering in the Semantic Web When/Where: 12:20--1:10 Friday November 6, 2009, room 306 Boyd Abstract: Searching the Internet has become a part of our lives. It is hard to find a right answer because of having a lot of information. Also, in order to get an answer from computers, they need to understand human language. Question Answering focusses on how to solve those problems to give us better answers to our questions. I will talk about what Question Answering is and two current approaches. -------------------------------------------------------------------------------- Speaker: Man Chon (Kevin) U Title: Nonlinear Forecasting for Financial Markets When/Where: 12:20--1:10 Friday October 23, 2009, room 306 Boyd Abstract: Researchers have known for some time that nonlinearity exists in the financial markets and that neural networks can be used to forecast market returns. In this presentation, we present a novel stock market prediction system which focuses on forecasting the relative tendency (percentage of change) between different stocks rather than purely predicting their values. This research utilizes artificial neural network models for estimation. The results are examined for their ability to provide an effective forecast of future values. Certain techniques, such as sliding windows and chaos theory, are employed for data preparation and preprocessing. Our system successfully predicted the relative tendency of different stocks with up to 57.64% accuracy. -------------------------------------------------------------------------------- Speaker: Rohit Mullangi Title: Scalable Decentralized Search for Peer-to-Peer Systems When/Where: 12:20--1:10 Friday October 16, 2009, room 306 Boyd Abstract: Napster pioneered the idea of peer-to-peer file sharing, and supported it with a centralized file search facility. Subsequent P2P systems like Gnutella adopted decentralized search algorithms. However, Gnutella's notoriously poor scaling led some to propose distributed hash table solutions to the wide-area file search problem. Contrary to that trend the authors of this paper advocate retaining Gnutella's simplicity while proposing new mechanisms that greatly improve its scalability. -------------------------------------------------------------------------------- Speaker: Douglas Brewer Title: Combating Click Fraud with Duplicate Detection When/Where: 12:20--1:10 Friday October 9, 2009, room 306 Boyd Abstract: The World Wide Web has exploded in popularity and is now used by an enormous number of people. People using the World Wide Web have come to expect that websites will provide them content for free. Thus, websites have come to rely on advertising as the primary means of financial support. The popular Pay-Per-Click method for showing advertisements, used by Google AdSense, Yahoo! Publisher Network, and Microsoft pubCenter, has led to a new form of online fraud termed Click Fraud. Click Fraud has become a grave concern, stealing as much as one billion dollars. I will discuss a new method for detecting Click Fraud called duplicate detection. Duplicate detection works by finding duplicate clicks on advertisements and removing them. Duplicate detection shows promising results with very low error rates detecting click fraud. -------------------------------------------------------------------------------- Speaker: Vasim Mahamuda Title: Computational Metagenomics -- Discovering the "Unculturable" Microbial World When/Where: 12:20--1:10 Friday October 2, 2009, room 306 Boyd Abstract: Metagenomics, the application of genome sequencing techniques to unculturable communities, is revolutionizing microbiology and has shed light on the role of microbial communities in our environment as well as human health. This relatively new field of genetic research enables studies of organisms that are not easily cultured in a laboratory. In my presentation I will focus on what Metagenomics means, a tutorial-style introduction to various technical terms involved in computational biology, the different sequencing techniques, and about how computational methods can contribute to this field. Towards the end I will talk about a real world problem of separating mixed DNA sequences of two different species using machine learning techniques. -------------------------------------------------------------------------------- Speakers: Rohit Mullangi and Charu Vijayagopal (half the period each) Titles: JavaFX (Mullangi) ROBOFORM Password manager (Vijayagopal) When/Where: 12:20--1:10 Friday September 25, 2009, room 306 Boyd Abstract (Mullangi): JavaFX is a rich client platform for building cross-device applications and content. It is designed to enable easy creation and deployment of rich Internet applications (RIAs) with immersive media and content. The JavaFX platform ensures that RIAs look and behave consistently across diverse form factors and devices. The JavaFX platform empowers content developers by enabling them to focus on creativity instead of coding. It enables developers to create game-changing applications and engaging content with maximum market penetration opportunities. Abstract (Vijayagopal): Roboform is an award winning password manager which also does provides the following services: 1. Encrypts passwords for total security 2. Fills out long online forms 3. Generates secure random passwords 4. It remembers the web address along with the user name and password. It thus provides one click easy login to many sites. 5. You can actually see your saved passwords and when you change them Roboform automatically updates its stored data. Using this makes managing multiple logins and password management safe and easier. All you have to do is just remember one master password.. JUST ONE password for Roboform. I will discuss some of its security features and how to use it well. -------------------------------------------------------------------------------- Speakers: Vasim Mahamuda and Jeremy Jones (half the period each) Titles: Cloud Computing (Mahamuda) Programming in Microsoft Visual Studio 2008 and XNA (Jones) When/Where: 12:20--1:10 Friday September 11, 2009, room 306 Boyd Abstract (Mahamuda): Cloud computing is a style of computing in which dynamically scalable and often virtualized resources are provided as a service over the Internet. Cloud computing refers to a set of services that provide companies and application developers with the means to scale their application's resource usage through the Internet on demand. In my talk, I will give an introduction of what is cloud computing, the need for cloud computing, services offered, advantages and the major companies involved in cloud computing. Abstract (Jones): The talk will cover a brief background on the two programs, basic concepts of using Visual Studio 2008, and how to use XNA to port games over to the XBOX 360 console. -------------------------------------------------------------------------------- Speakers: Kitae Myoung and Qi Li (half the period each) Titles: Dreamweaver (Myoung) and Apache Ant (Li) When/Where: 12:20--1:10 Friday September 4, 2009, room 306 Boyd Abstract (Myoung): I will be focused on how to use Dreamweaver, but in the beginning I will talk about what is Dreamweaver and compare it to notepad and talk about its advantage and disadvantage. I will show the students how to make a webpage on a web server. During this work, I will introduce basic useful icons and explain what we have to do for making webpages. Abstract (Li): Apache Ant is a Java-based build tool, a very popular program designed to help software teams by automating all the processes of compiling code, running tests, and packaging the results for redistribution. Ant is written in Java and uses XML syntax as the configuration files. Ant is created to be cross-platform, easy to use, extensible, and scalable. Ant can be used not only in a small personal project but also in a large, multiteam software project. Ant aims to automate your entire build process. In my talk, I will give an overview of Ant and demonstrate how Ant works. -------------------------------------------------------------------------------- Speakers: Douglas Brewer and Man Chon (Kevin) U (half the period each) Titles: The Apache Filter API (Brewer) User-Oriented Document Summarization through Vision-Based Eye-Tracking (Kevin) When/Where: 12:20--1:10 Friday August 28, 2009, room 306 Boyd Abstract (Brewer): Apache is one of the world's most widely used web servers. Its primary purpose is to serve static HTML documents, but through the use of its module system, Apache can server dynamic content. The most well known module using the Apache Filter API is PHP, a server-side scripting language for generating dynamic content. In my talk, I give an overview of the Apache Filter API and a working example of a filter modifying an HTML document. Abstract (Kevin): In our current information era, there is an explosion of reference materials for professionals working in nearly every area. Therefore, the need for automatic document summarization has become compelling. Automatic summarization has continuously been an active research area in artificial intelligence, information retrieval and natural language processing for more than two decades. However, up till now, studies on document summarization have rarely taken human factors into consideration, which means that existing automatic summarization methods have not emphasized the reading interest of individual readers. To address this problem, the authors of this paper propose a new document summarization algorithm which is personalized. Link to the paper: http://portal.acm.org/citation.cfm?id=1502650.1502656 --------------------------------------------------------------------------------