2001-2002 AbstractsCPSC 681 Graduate SeminarNext-Generation Information NetworkingAvi Silberschatz, Bell Laboratories
4:10pm, Monday January 14, 2002
AbstractNext generation information networking will incorporate the best features of today's voice and data networks. They will seamlessly combine communications, software, and contents. These systems will need to deal with multimedia data, the Web, and real-time applications. Some of the key challenges that need to be addressed are the need for handling exabyte of data in terms of storage and delivery, quality of service, availability, security, and privacy.This talk will present a grand tour of the wide variety of issues facing next-generation information networking, highlight their characteristics, and introduce a few research projects carried out at Bell Labs that address these challenges. We conclude the talk with seven predictions for the new millennium.
BiographyAvi Silberschatz is the Vice President of the Information Sciences Research Center at Bell Laboratories, Murray Hill, New Jersey. Prior to joining Bell Labs, he held a chaired professorship in the Department of Computer Sciences at the University of Texas at Austin. His research interests include operating systems, database systems, real-time systems, storage systems, and distributed systems.In addition to his academic and industrial positions, Silberschatz served as a member of the Biodiversity and Ecosystems Panel on President Clinton's Committee of Advisors on Science and Technology, as an advisor for the National Science Foundation, and as a consultant for several private industry companies. Professor Silberschatz is an ACM Fellow and an IEEE Fellow. He received the 1998 ACM Karl V. Karlstrom Outstanding Educator Award, the 1997 ACM SIGMOD Contribution Award, and the IEEE Computer Society Outstanding Paper award for the article "Capability Manager," which appeared in the IEEE Transactions on Software Engineering. His writings have appeared in numerous ACM and IEEE publications and other professional conferences and journals. He is a co-author of two well known textbooks -- Operating System Concepts and Database System Concepts. Facutly Contact: John Leggett (leggett@cs.tamu.edu) CPSC 681 Graduate Seminar:UrbanSim: Integrated Land Use, Transportation and Environmental ModelingAlan Borning, University of Washington
4:10pm, Wednesday January 23, 2002
AbstractPatterns of land use and available transportation systems play a critical role in determining the economic vitality, livability, and sustainability of urban areas. Transportation interacts with land use, and both have significant environmental effects, in particular on emissions, resource consumption, and conversion of rural to urban land. Decisions concerning land use and transportation are frequently contentious; further, the long-term consequences of decisions may not be apparent at the time they are being made. Strong technical support may help foster informed deliberation on these issues. Toward this goal, we have been developing UrbanSim, a reusable urban land use and transportation modeling system. Our purpose is to provide a tool for citizens' groups, urban planners, elected officials, and others to help predict future patterns of urban development and impact under different possible scenarios over periods of 20 or more years.The talk will include an overview of the modeling activity, followed by a technical discussion of the system and of particular computational aspects, including designing a software architecture to support relatively independent implementation and evolution of the different component models, automatically choosing appropriate visualizations, and supporting new model development with a domain-specific programming language. I'll conclude with a discussion of current and future research directions. In human computer interaction, these include providing more effective ways of understanding the results from and interacting with complex simulations, and ways of linking stakeholder values with design choices in simulations and their interfaces. In software engineering, research directions include supporting fully disaggregated microsimulation, including programming language support and improved techniques for semantically-based implicit invocation. In computer graphics, we plan to produce simulated street-level animations of urban environments from a policy-driven simulation; and in statistics, we want to apply Bayesian networks and multi-agent microsimulation to support complex inference modeling in a problem domain with inherent uncertainty. This is joint work with Paul Waddell and others at the University of Washington.
BiographyAlan Borning is a professor in the Department of Computer Science & Engineering at the University of Washington, Seattle, USA, where he has been since 1980. His research interests are in constraint-based languages and systems, human computer interaction, and integrated land use, transportation, and environmental modeling. He received a BA degree from Reed College in 1971, and a PhD from Stanford University in 1979.Facutly Contact: Frank Shipman (shipman@cs.tamu.edu) CPSC 681 Graduate SeminarEvolutionary Algorithm Based Off-line / On-line Path Planner for UAV NavigationKimon Valavanis, Technical University of Crete
4:10pm, Wednesday January 30, 2002 AbstractAn Evolutionary Algorithm based framework, a combination of modified Breeder Genetic Algorithms incorporating characteristics of classic Genetic Algorithms, is utilized to design an off-line / on-line path planner for Unmanned Aerial Vehicles (UAVs) autonomous navigation. The path planner calculates a curved path line with desired characteristics in a 3-D rough terrain environment, represented using B-Spline curves, with the coordinates of its control points being the Evolutionary Algorithm artificial chromosome genes. Given a 3-D rough environment and assuming flight envelope restrictions, two problems are solved: i) UAV navigation using an off-line planner in a known environment, and, ii) UAV navigation using an on-line planner in a completely unknown environment. The off-line planner produces a single B-spline curve that connects the starting and target points with a predefined initial direction. The on-line planner, which is based on the off-line one, given on-board radar readings, gradually produces a smooth 3-D trajectory aiming at reaching a predetermined target in an unknown environment; the produced trajectory consists of smaller B-Spline curves smoothly connected with each other. Both planners have been tested under different scenarios and they have been proven effective in guiding an UAV to its final destination, providing near-optimal curved paths quickly and efficiently. CPSC 681 Graduate SeminarHow to Write a Research Paper, Thesis, or Dissertation: Scholarly Writing as StorytellingFrank Shipman , Texas A&M University
4:10pm, Monday February 11, 2002
AbstractMany times, writing a research paper, thesis or dissertation seems to be an afterthought in the research process. Indeed, many research reports seem to be no more than documentation of the effort, laundry lists of tasks and facts, without making an effort to communicate effectively. This presentation talks about why writing the report should be viewed as intertwined within the research process and how the scholarly writing parallels storytelling. This is a revised version of a presentation I give at doctoral symposiums for advanced Ph.D. students getting ready to write their dissertations. BiographyFrank M. Shipman III is an Associate Professor in the Department of Computer Science and Center for the Study of Digital Libraries at Texas A&M University. He has been pursuing research in the areas of hypermedia, computer-supported cooperative work, and intelligent user interfaces since 1987. Dr. Shipman's doctoral work at the University of Colorado and subsequent work at Xerox PARC and Texas A&M has investigated combining informal and formal representations in interfaces and methods for supporting incremental formalization. Dr. Shipman's research has led to commercial systems like ForeFront Group's Virtual Notebook System and Eastgate System's Web Squirrel. He manages two on-going research projects in the areas of spatial hypertext and computers and education. CPSC 681 Graduate Seminar1 Million Monkeys Banging on Keyboards - Is there any theory behind GUI testing?Bob Metzger, Hewlett-Packard
4:10pm, Monday February 18, 2002
AbstractMost commercial applications that have Graphical User Interfaces are tested manually, with drudge labor typing, mousing, and eyeballing screens. There are almost no papers published that describe a systematic approach to automatic testing of GUI testing. This talk presents a method that applies some simple graph theory concepts and the latest design approaches to GUI software to create a method for systematically generating GUI test cases that can be automatically executed and checked. BiographyBob Metzger has worked in software development for over 25 years, including work at I.P.Sharp Associates and Convex Computer Corp. His work at Convex produced the first commercially available compiler that performed interprocedural optimizations on multiple languages. He was also awarded a $443,000 grant from DARPA resulting in his textbook: Automatic Algorithm Recognition and Replacement: A New Approach to Program Optimization, co-authored with Zhaofang Wen, published by MIT Press, 2000. He is currently a Senior Software Engineer at Hewlett Packard, developing new technologies for analyzing the performance of applications. CPSC 681 Graduate SeminarDNA Computing on SurfacesAnne Condon , University of British Columbia
4:10pm, Monday February 25, 2002
AbstractIn 1994, Len Adleman described how to solve a small instance of a famous combinatorial problem - the Traveling Salesman problem - in a novel way. Adleman's method was to efficiently create a test tube of DNA strands, each representing a possible solution to the problem, and to extract the elusive true solution using tools from molecular biology. Adleman's work raises many questions at the interface of biochemistry, mathematics, and computer science. How can information be efficiently and reliably stored in, and subsequently retrieved from, DNA molecules? How can logical operations be performed on information-carrying DNA strands? Can new nanostructures and materials be assembled in a programmable fashion from DNA? What might be useful applications of DNA computing? In this talk, we describe our collaborative work that addresses these questions. The premise of the Wisconsin DNA computing group is that solid-phase chemistry will be a critical technology in realizing DNA computation, because it allows a much greater degree of control in the chemical processes than that achievable via solution-phase methodologies. With this approach, many DNA strands are immobilized on a planar surface, and logical operations are performed on all of the strands in parallel, using chemical and enzymatic processes. The talk will describe some combinatorial problems that arise in building our prototype DNA computer, as well as results of a DNA computation for a small instance of the NP-hard Satisfiability problem. This is joint work with Professors Rob Corn and Lloyd Smith of the Chemistry Department at U. Wisconsin-Madison, along with several students. BiographyAnne Condon's research contributions are in the areas of probabilistic and interactive complexity classes, design and analysis of algorithms for computationally intractable problems, and DNA computing. Condon received a B.Sc. degree from University College, Cork, Ireland in 1982 and a Ph.D. from the University of Washington in 1987. Her thesis, a study of game-theoretic complexity classes, won an ACM Distinguished Dissertation Award in 1988. Condon received a National Young Investigator Award in 1992. She was named Distinguished Alumna of University College Cork in 2001, for her contributions in the area of DNA computing. Condon is currently a Professor in the Department of Computer Science at U. British Columbia. She was a faculty member of the Computer Sciences Department at U. Wisconsin between 1987 and 1999. CPSC 681 Graduate SeminarGenesis: A Distributed Data Collection and Simulation System for Network Management and ProtectionDr. Boleslaw Szymanski , Rensselaer Polytechnic Institute
4:10pm, Wednesday March 20, 2002
AbstractGenesis proposes a novel approach to scalability and efficiency of parallel network monitoring, modeling and simulation for network management and intrusion detection. The basis of our approach is network decomposition that creates separate network domains. Each domain is independently monitored, modeled and simulated by separate software components. Domain monitoring involves a repository of data and collection agents while models and simulations are run by dedicated processor clusters allocated to each domain. Repositories and simulators collaborate on exchanging data and models to keep the network simulations consistent between domains. One of the desirable features of our design is its independence from the underlying simulators and repositories running in the individual domains. Hence, our architecture can be integrated with a number of existing technologies thereby supporting system interoperability. Our primary application is network management. The simulation in this application predicts changes in the network performance caused by network parameter tuning. Growing security threats to networks increase the importance of intrusion and anomaly detection. Using data collection agents, we create unique signatures for attacks or legitimate user activities (monitoring a stream of network packets at the server). Signatures are represented as probabilistic or time-dependent finite state automata and their parsers can easily be embedded into mobile agents. BiographyDr. Boleslaw Szymanski is currently Associate Dean for Information Technology and Professor of Computer Science at Rensselaer Polytechnic Institute in Troy, New York. His research interests include network management, modeling and simulation, network middleware and distributed and Web-based computing, scientific parallel computation, algorithm design and verification for parallel and distributed systems, and modeling and simulation of biological and ecological systems. CPSC 681 Graduate SeminarOn Fighting Two Varieties of SpamCynthia Dwork, Microsoft Research, Silcon Valley Campus
4:10pm, Monday March 25, 2002
AbstractThe internet and the web are public places with concommitant exposure to obnoxious behavior. In this talk we discuss two forms of "spam," junk e-mail and search engine spam, and present two different technologies for fighting back. BiographyCynthia Dwork received her PhD from Cornell University in 1983 under the supervision of John Hopcroft. After a two-year post-doc at MIT, she joined the IBM Almaden Research Center, where she remained until becoming Compaq Staff Fellow in 2000. In 2001 she joined the nascent Silicon Valley campus of Microsoft Research. Most of Dwork's research has been in cryptography and other topics in distributed computing. She is the co-inventor of non-malleable cryptography and of the only public-key cryptosystem for which random instances are provably as hard to break as the hardest instances of the underlying mathematical problem. Faculty Contact: Jennifer Welch (welch@cs.tamu.edu)CPSC 681 Graduate SeminarDetecting Traffic Patterns at High Speeds in RoutersGeorge Varghese, UC San Diego
4:10pm, Wednesday March 27, 2002
AbstractIn this talk I will describe my research into detecting traffic patterns in real-time in Internet Routers. Any such processing must be done in a packet arrival time (8 nsec at the highest link speeds today) and hence must take a small number of memory references, and also store state in high speed memories (analogous to cache or register memory) that are limited in size. The problem is to detect important patterns (e.g., Internet lookups, Denial-of-Service Attacks) in a small constant number of operations using a relatively modest amount of state. I will first briefly survey the work our group has done on detecting both patterns *within* packets (e.g., IP lookups, firewall filters) as well as patterns *between* packets (e.g., Quality of Service). I will then describe in detail two larger examples. The first is a new scalable packet classification scheme (SIGCOMM 2001) and new scalable measurement schemes (IMW 2001). The scalable measurement algorithms show how to quickly process a traffic stream to identify the customers that send more than a threshold (e.g., 1% of the total bandwidth) using memory only proportional to the maximum number of such customers (e.g., 100). This in turn makes possible a new form of accounting called threshold accounting in which only flows above a threshold are charged by usage, while the rest are charged a fixed fee. Threshold accounting generalizes the standard notions of usage-based and duration-based pricing. BiographyGeorge Varghese worked at DEC for several years designing DECNET protocols before obtaining his Ph.D in 1992 from MIT. He joined Washington University in 1993 as an Associate Professor where he won the ONR Young Investigator Award in 1996. He is currently a Professor at the University of California, San Diego where he works on efficient protocol implementation and protocol design. Several of the algorithms he has helped develop (e.g., IP Lookups, timing wheels, DRR) have found their way into commercial systems that range from HotMail to the Cisco GSR Router. Faculty Contact: Jennifer Welch (welch@cs.tamu.edu)CPSC 681 Graduate SeminarComputational Understanding of Gene Regulation: From Gene Expression to Gene NetworksVladimir Filkov, State University of New York at Stony Brook
4:10pm, Monday April 1, 2002
AbstractAn area of very active research in Computational Biology is gene regulation and gene regulatory networks. Genes and gene products regulate all of the processes in the cell, as they react to developmental or environmental events. Active (expressed) genes regulate the activity (expression) of other genes by coding for proteins that physically interact with their DNA. In that sense, all the genes in an organism are parts of a gene regulatory network that describes the activities and relationships of all the genes of that organism. The therapeutic benefits of having a blueprint of a gene network are enormous because the organism's responses can then be understood and even modified. In this talk we present our past and current work on inference of gene regulation through computational analysis of gene expression data. We describe (1) one of the first software systems for gene network inference from gene expression data; (2) methods for analysis of gene expression data across experiments with application to gene regulation; and (3) a system we are building for integrated analysis of heterogeneous expression data. BiographyVladimir Filkov is a bioinformatics faculty candidate. He received his BS suma cum laude in Applied Mathematics/Computer Science from La Roche College, Pittsburgh PA in 1997. Subsequently, Mr. Filkov completed a Master's degree in Computer Science at State University of New York at Stony Brook in 1999. He is currently working on his PhD under supervision of Prof. Steven Skiena, and his expected graduation date is May 2002. His research interests are computational biology, combinatorial algorithms and optimization, network inference, and biological data mining. CPSC 681 Graduate SeminarPhysically-Based Modeling of Fire and CloudsJohn Keyser, CS Department, Texas A&M University
4:10pm, Wednesday April 10, 2002
AbstractModeling of physical phenomena is an important part of many computer graphics applications. While high quality images of such phenomena can be created with significant artistic input, in many instances the simulation of these processes is too slow to be useful for all applications. This is particularly true for interactive simulations. Recent research in the graphics community has resulted in stable simulators for fluid motion. These simulators achieve significantly higher speeds than previous simulators, while maintaining a high degree of visual fidelity. In this talk, I will discuss the recent work we have performed on simulation of flames and clouds. I will describe simple models for flame formation and spread and cloud generation and motion. These models make use of the fast fluid simulators to achieve interactive models of flames and clouds. The work I will present has been the focus of work by two of my graduate students, Zeki Melek and Derek Overby, over the past year. BiographyJohn Keyser is an Assistant Professor in Computer Science at Texas A&M University. He received his Ph.D. in 2000 from the University of North Carolina at Chapel Hill. His research interests are broadly in the areas of graphics and computer algebra, with a particular emphasis on geometric modeling and computation. CPSC 681 Graduate SeminarCombining Approaches to Finding Motifs in DNA SequencesSing-Hoi Sze, Department of Computer Science and Engineering, University of California at San Diego
4:10pm, Monday April 15, 2002
AbstractMotif finding (pattern discovery in unaligned DNA sequences) is a fundamental problem in both computer science and molecular biology with important applications in locating regulatory sites and drug target identification. We describe two combinatorial approaches WINNOWER and SP-STAR which prove to give better performance than some of the most popular signal finding algorithms, including CONSENSUS, GibbsDNA and MEME, when applied to simulated samples with uniform background distribution. We benchmark these and other algorithms on several bacterial and archaeal sites in a setting specifically designed to imitate the situations arising in comparative genomics studies. We then describe a few collaborative projects with biologists to apply these motif finding algorithms to real biological problems, including finding new Drosophila immunity genes and finding potential sites cooperating with NRSE sites in neuronal genes. Finally, we describe an ongoing effort to extend SP-STAR to model dyad motifs which consist of two parts with varying distances between the parts. BiographySing-Hoi Sze is currently a postdoctoral research associate at the University of California at San Diego working on algorithms for computational biology problems. He received his Ph.D. in computer science from the University of Southern California in December 2000. His thesis was on the development of combinatorial approaches to gene finding and signal finding in DNA sequences. His current research areas include motif finding algorithms and their applications, and algorithms for EST sequence assembly and enumeration of alternatively spliced variants of a gene. CPSC 681 Graduate SeminarGraduate Student Orientation (Aug 27, Aug 29, Sept 3)Grad Orientation I: Key Departmental Rules & RegulationsCSGSA Officers, Bart Childs, Graduate Advisor, Key Departmental Staff, Texas A&M University
4:30pm, Monday August 27, 2001
AbstractStaff members in the department will help new students get familiar with the different offices in the Computer Science Department. The Graduate Advisor will go over some details related to the academic program. Other staff members will discuss computational resources (accounts, printing, webpages, etc), office stuff (getting an office, a mailbox, etc), and many other useful things. (It's not a bad refresher for current students as well!) CPSC 681 Graduate SeminarGrad Orientation I: Continued - Academic Program DetailsBart Childs, Graduate Advisor Texas A&M University
4:20pm, Wednesday August 29, 2001
AbstractThe Graduate Advisor will go over some details related to the academic program that were not covered on Monday, such as the new 30 hours MCS program. CPSC 681 Graduate SeminarGrad Orientation II: Faculty, Research Opportunities, and Pizza!CS Faculty, Texas A&M University
4:10pm, Monday September 3, 2001
AbstractFaculty members in the department will present their research. Followed by this, there is a poster presentation by the students. The posters will be in room 126 and you can enjoy Pizza and sodas as you go around to see the posters and chat with the faculty and students. CPSC 681 Graduate SeminarParasitic Extraction of VLSI CircuitsVivek Sarin, CS Department, Texas A&M University
4:10pm, Wednesday September 12, 2001
AbstractIn the pursuit for faster microprocessors, the semiconductor industry is increasingly using deep sub-micron technology to design VLSI circuits. At very high frequencies, signal propogation across these circuits is delayed considerably by the "parasitic" effects of on-chip capacitance and inductance due to the interconnect. The ability to estimate the resistive, capacitive, and inductive (RCL) parasitic quickly and accurately is critical to the design and verification of large VLSI circuits. The main challenge to such an effort is the design of algorithms to solve very large and dense complex linear systems of equations. To estimate the RCL parasitic for realistic circuits, one must exploit parallelism in these algorithms by using high-performance software on parallel computers. This talk will present novel algorithms for computing the parasitic of VLSI circuits. These algorithms have superior convergence properties as well as high degree of concurrency that can be exploited to develop efficient implementations on parallel architectures. BiographyDr. Sarin's research interests include numerical algorithms, parallel algorithms, and their application to scientific and engineering disciplines. He has taught graduate courses on numerical methods, computational linear algebra, and parallel and distributed Numerical algorithms. Dr. Sarin received his Ph.D. in Computer Science from University of Illinois at Urbana-Champaign in 1997. He joined Texas A&M in Fall 1999 after spending two years as a research associate at Purdue University. CPSC 681 Graduate SeminarHardware-Software Codesign: Issues and ChallengesRabi Mahapatra, CS Department, Texas A&M University
4:10pm, Monday September 17, 2001
AbstractThe increasing complexity in recent embedded systems can no longer follow the traditional design steps due to their specification constraints and reduced time to market strategy. Thus, the hardware-software codesign approach has been introduced to meet these requirements. In codesign, the hardware and software design steps are considered simultaneously starting from the system specification to their implementation. In this introductory talk, we shall review the existing codesign framework and discuss design challenges associated at various steps. Issues such as partitioning, cosimulation and design space exploration will be outlined. The CPU cost leverage by performance is increasingly limited by power and cooling. Current efforts on power modeling and estimation for embedded systems will be presented. BiographyDr. Rabi Mahapatra has research interests in system architecture and parallel processing. Currently, he is pursuing research in the area of Embedded System Codesign. Dr. Mahapatra received his Ph.D. from Indian Institute of Technology in 1992. He is currently an Associate Professor at Texas A&M. Before joining at A&M, he was an Assistant Professor in ECE department at Indian Institute of Technology, Kharagpur. He is a Senior Member of IEEE Computer Society. CPSC 681 Graduate SeminarReinforcement Learning in Non-Stationary Environments and Dynamic Network Routing AlgorithmsDevika Subramanian, Rice University
4:10pm, Wednesday September 26, 2001
AbstractI will present two new distributed learning algorithms for network routing based on reinforcement learning in non-stationary environments. The first algorithm uses "ants" to explore the network and learn good routes. The "ant" algorithm has very good resilience properties; random corruption of router state has limited impact on path computation. When failure rates in the network are high, the algorithm outperforms traditional packet routing schemes. However, the algorithm is not competetive with traditional methods under more realistic conditions. An extensive experimental analysis of the ant algorithm led to the design of a new hybrid algorithm in which destination-initiated ant-style path learning is added on top of a traditional routing algorithm. This hybrid algorithm provides shortest paths to "hot" destinations in a network with respect to a dynamic (traffic-sensitive) cost metric. It is one of the first in the routing literature to provide stable routing with dynamic metrics; it relies on independent, time-staggered recomputation of paths to the destinations receiving the most traffic. Experimental results on a realistic network simulator on Internet-like topologies suggest that the method offers significant throughput gains over traditional static metric routing, with very minimal overheads. This work, which was funded in part by SW Bell, is joint with Professor Peter Druschel and Johnny Chen of Rice University. BiographyDevika Subramanian obtained her undergraduate degree in electrical engineering and computer science from the Indian Institute of Technology, and her PhD in computer science from Stanford University in 1989. Her research interests are in the design and analysis of embedded adaptive systems and their applications in science and engineering (http://www.cs.rice.edu/~devika). She was a faculty member in Computer Science at Cornell University before she joined Rice in 1995. Her research is currently funded by NSF, ARPA, NASA, ONR and Texas ATP. CPSC 681 Graduate SeminarVisual Perceptual Grouping in a Self-Organizing Map of Spiking NeuronsYoonsuck Choe, CS Department, Texas A&M University
4:10pm, Monday October 1, 2001
AbstractPerceptual grouping is the process of identifying the constituents in the visual scene that together form a coherent object. The goal of my research is to understand the neural mechanisms of perceptual grouping. The hypotheses are that (1) perceptual grouping is carried out through synchronized firing of neurons representing the same object, and that (2) self-organized lateral connections encoding statistical regularities of the visual environment mediate such a synchronization. A self-organizing neural network of spiking neurons was developed to test these hypotheses in the perceptual grouping task of contour integration. The network self-organized orientation maps and patchy lateral connections similar to those found in the visual cortex, and the contour integration, segmentation, and completion performance measured by the degree of synchrony in neural populations accurately predicted human performance. Such results suggest that synchronized activity can represent perceptual events, and statistical properties of the input can shape the structure of the cortex and the perceptual performance. By providing a computational framework where image statistics, perceptual performance, and neural structure can be compared, the model helps us understand the neural mechanisms of perceptual grouping, and enable us to build better computational vision systems to tackle real-world vision problems. BiographyYoonsuck Choe is an Assistant Professor of Computer Science at Texas A&M University. He received a B.S. in Computer Science from Yonsei University, Korea, in 1993, and an M.A. and a Ph.D. in Computer Sciences from the University of Texas at Austin in 1995 and 2001. His current research interests include modeling of the visual cortical structure and function with neural networks, understanding the role of statistical properties in natural images in shaping the structure of the cortex, applications in computer vision, and pattern recognition. CPSC 681 Graduate SeminarDemonstrational Customization of a Shared Whiteboard to Support User-Defined Semantic Relationships among ObjectsDu Li , CS Department, Texas A&M University
4:10pm, Wednesday October 10, 2001
AbstractAs a promising approach to end-user computing, programming by demonstration (PBD) techniques have been explored by many researchers in single-user applications. This paper investigates how PBD techniques can be used to support end-user customization of groupware tools. In collaborative applications, complex semantic relationships can emerge unanticipatedly among objects (participants, data artifacts, tools, devices, etc.) such as the hierarchical organization of participants, consistency maintenance among different views of the same data, and awareness control. It is important that end users are empowered to customize the groupware application to model and enforce such relationships. We present an active rule based approach to modeling user-defined semantic relationships in collaborative applications and explore a demonstrational approach for end-user customization of collaboration tools to support the definition of those relationships. The presented work is based on our work on a shared whiteboard tool, CAB or Collaborative Active whiteBoard. Our approach is being generalized to support end-user defined object relationships in shared workspaces. This work will be presented at ACM GROUP'01 Conference on Supporting Group Work, Sep.30-Oct.3, Boulder, Colorado. A full version of paper is available from the speaker's web page http://www.csdl.tamu.edu/~lidu. BiographyDr. Du Li received his PhD degree from UCLA in June 2000. Dr. Li joined Texas A&M as an assistant professor in July 2000. His research interests include CSCW, Distributed Systems, Middleware, Programming Languages, Logic Programming, and Databases. His current research focuses on collaboration modeling and infrastructure support. CPSC 681 Graduate SeminarEmergence and Evolution of Spatial Hypertext and the Visual Knowledge BuilderFrank Shipman , CS Department, Texas A&M University
4:10pm, Monday October 15, 2001
AbstractThe development of spatial hypertext systems has been driven by the need to lower the user's effort of expression when organizing information. Traditional hypertext, such as the web, requires authors to create explicit representations of relations between information nodes in the form of links. Spatial hypertext lets users express categories and relationships through visual similarity and co-location of information objects. We have developed the Visual Knowledge Builder (VKB), a spatial hypertext system that also records the process of authoring in real time. Three years of experience using VKB for note taking, project and conference management, and scholarly writing underscores the value of supporting lightweight interpretation in information systems. BiographyFrank M. Shipman III is an Associate Professor in the Department of Computer Science and Center for the Study of Digital Libraries at Texas A&M University. He has been pursuing research in the areas of hypermedia, computer-supported cooperative work, and intelligent user interfaces since 1987. Dr. Shipman's doctoral work at the University of Colorado and subsequent work at Xerox PARC and Texas A&M has investigated combining informal and formal representations in interfaces and methods for supporting incremental formalization. Dr. Shipman's research has led to commercial systems like ForeFront Group's Virtual Notebook System and Eastgate System's Web Squirrel. He manages on-going research projects in the areas of spatial hypertext, information visualization, requirements gathering, and computers and education. CPSC 681 Graduate SeminarSocial Creativity and Meta-Design in Lifelong Learning CommunitiesGerhard Fischer, University of Colorado at Boulder
4:10pm, Monday October 22, 2001
AbstractComplex design problems require more knowledge than any one single person can possess, and the knowledge relevant to a problem is often distributed and controversial. Rather than being a limiting factor, the symmetry of ignorance (or asymmetry of knowledge) can provide the foundation for social creativity. Bringing different points of view together and trying to create a shared understanding among all stakeholders can lead to new insights, new ideas, and new artifacts. Social creativity can be supported by new media that allow owners of problems to contribute to framing and solving these problems. These new media need to be designed from a meta-design perspective by creating environments in which stakeholders can act as designers and be more than consumers. BiographyGerhard Fischer is a professor of Computer Science, a fellow of the Institute of Cognitive Science, and the director of the Center for LifeLong Learning & Design (L3D) at the University of Colorado at Boulder. Current research interests include new media supporting lifelong learning, human-human and human-computer collaboration, (software) design, domain-oriented design environments and universal design (assistive technologies). More information about the (L3D) center can be found at: http://www.cs.colorado.edu/~l3d/ . Facutly Contact: Frank Shipman (shipman@cs.tamu.edu)CPSC 681 Graduate SeminarScalable peer-to-peer substrates: A new foundation for distributed applicationsPeter Druschel, Rice University
4:10pm, Monday November 5, 2001
AbstractPeer-to-peer (p2p), initially conceived for the purpose of sharing music in the Internet, proves itself to be a more general paradigm for organizing large-scale distributed applications. We define p2p systems broadly as self-organizing, decentralized, distributed systems where most or all communication is symmetric. The self-organization, decentralization, diversity and numerocity of resources inherent in the approach lend themselves to a large domain of applications beyond file sharing, anonymity and anti-censorship. Conversely, the decentralization and diversity also pose difficult problems, particularly in resource management and security. Recent work on p2p overlay networks like CAN, Chord, Pastry and Tapestry has made significant strides towards providing a general substrate that simplifies the construction of a wide range of p2p applications. These overlay networks effectively shield applications from the complexities of organizing and maintaining an overlay network, and from distributing and locating resources. In this talk, I'll present an overview of Pastry, a p2p overlay network that provides self-organization, efficient resource location and distribution, and incorporates interesting heuristics that exploit proximity in the underlying Internet. I will also sketch two applications built upon Pastry to date: PAST, an archival, cooperative file storage and distribution facility, and SCRIBE, a highly scalable event notification system. I'll conclude with an outlook on key research problems and future directions. BiographyPeter Druschel is an Associate Professor of Computer Science at Rice University. He received the Dipl.-Ing. (FH) degree from Fachhochschule Muenchen, Germany, in 1986, and the Ph.D. degree from the University of Arizona in 1994, under the direction of Larry L. Peterson. He received an NSF CAREER Award in 1995 and a Alfred P. Sloan Fellowship in 2000. He is a member of the Computer Systems Laboratory. During the 2000-2001 academic year, Peter was on sabbatical leave, which he spent with the SRC group at Laboratoire d'Informatique de Paris 6 (LIP6) (May-June 2000), the Cambridge Distributed Systems group at Microsoft Research Cambridge, UK (August-December 2000), and the PDOS group at the MIT Laboratory for Computer Science (January-June 2001). CPSC 681 Graduate SeminarRandomized Motion Planning: From Intelligent CAD to Protein Folding Nancy Amato, CS Department, Texas A&M University
4:10pm, Wednesday November 7, 2001
AbstractMotion planning arises not only in robotics but in many other areas such as intelligent CAD (virtual prototyping), mixed reality systems (training and computer-assisted operation), and even computational biology and chemistry (protein folding and drug design). Surprisingly, a single class of planners, called probabilistic roadmap methods (PRMs), have proven effective on problems from all these domains. Strengths of PRMs, in addition to versatility, are simplicity and efficiency, even in high-dimensional configuration spaces. Moreover, PRMs are (almost) embarrassingly parallel. In this talk, we introduce the PRM framework and give an overview of several PRM variants developed in our group to address the `narrow passage problem' (when the solution path must pass through confined regions which is a case not handled effectively by initial PRMs). We present results for difficult problems typical of virtual prototyping, and show that in some cases a hybrid system incorporating both an automatic planner and haptic user input leads to superior results. We will also describe our recent application of PRMs to protein folding. Currently we assume we know the native fold, and use our PRM-based technique to construct folding pathways from arbitrary starting conformations to the native fold. We have validated our results by comparing the secondary structure formation on our paths with known experimental results. More information regarding our work, including movies, can be found at http://www.cs.tamu.edu/faculty/amato/dsmft. BiographyNancy M. Amato joined Texas A&M University in January 1995, where she has been an associate professor since September 2000. She received B.S. and A.B. degrees in Mathematical Sciences and Economics, respectively, from Stanford University, and M.S. and Ph.D. degrees in Computer Science from UC Berkeley and the University of Illinois at Urbana-Champaign. Her main areas of research focus are motion planning, high-performance computing, and computational geometry. CPSC 681 Graduate SeminarBringing Supercomputing to Everyday LifeChekib Akrout, IBM Austin 4:10pm, Monday November 12, 2001
Abstract
Biography
CPSC 681 Graduate SeminarDistributed Token Circulation on Mobile Ad Hoc NetworksJennifer Welch , CS Department, Texas A&M University
4:10pm, Monday November 19, 2001
AbstractThis paper presents several distributed algorithms that cause a token to continually circulate through all the nodes of a mobile ad hoc network. An important application of such algorithms is to ensure total order of message delivery in a group communication service. Some of the proposed algorithms are aware of, and adapt to changes in, the ad hoc network topology. When using a token circulation algorithm, a round is said to complete when every node has been visited at least once. Criteria for comparing the algorithms include the average time required to complete a round, number of bytes sent per round, and number of nodes visited per round. Comparison between the proposed algorithms is performed using simulation results obtained from a detailed simulation model (with ns-2 simulator). This is joint work with Navneet Malpani and Nitin Vaidya. BiographyJennifer L. Welch received her B.A. degree (Liberal Arts) from the University of Texas at Austin in 1979, and her S.M. and Ph.D. degrees (Computer Science) from the Massachusetts Institute of Technology in 1984 and 1988. After spending one year at GTE Laboratories Incorporated in Waltham, MA and three years on the faculty at the University of North Carolina, Chapel Hill, she joined the Department of Computer Science at Texas A&M University, where she is currently an associate professor and interim department head. Her research interests include models, algorithms and lower bounds for distributed computing, with current emphasis on mobile ad hoc networks and randomized distributed data structures. CPSC 681 Graduate SeminarSolving Sequential Decision Tasks with NeuroevolutionRisto Miikkulainen, University of Texas at Austin
4:10pm, Monday November 26, 2001
AbstractIn many real world tasks such as game playing, resource management, and robot control, performance can be evaluated only after several decisions (or actions) have been made. This makes it difficult to come up with efficient decision strategies. Recently a promising approach has emerged: the strategy can be learned through genetic algorithms and represented in a neural network. I will describe an efficient implementation of this idea, and demonstrate it in a number of applications, including game playing and robot control. BiographyRisto Miikkulainen is an Associate Professor of Computer Sciences at the University of Texas at Austin. He received an M.S. in Applied Mathematics from the Helsinki University of Technology, Finland, in 1986, and a Ph.D. in Computer Science from UCLA in 1990. His current research includes models of sentence processing, self-organization of the visual cortex, episodic memory, and evolving neural networks with genetic algorithms. CPSC 681 Graduate SeminarA Microprocessor Design for 2014Doug Burger, University of Texas at Austin
4:10pm, Wednesday November 28, 2001
AbstractSilicon devices have at least 15 years of physical scaling left. It is unclear, however, that we will be able to exploit the properties and densities of those devices effectively. The one-two punch of clocking limits and slower wires will make performance scaling more difficult than ever before, as will the growing unreliability of these devices. In this talk, I will present a new class of microprocessor architectures, called Grid Processor Architectures (GPAs). GPAs are intended to show an order of magnitude higher performance, scale with technology, adapt to numerous application classes, and tolerate radiation or noise-induced soft errors. Finally, I will describe the prototype TRIPS chip being designed at UT-Austin, which contains multiple GPA cores on a single die.
BiographyDoug Burger is an Assistant Professor in Computer Sciences at the University of Texas at Austin. He received his Ph.D. in Computer Sciences from the University of Wisconsin-Madison in December, 1998. His main research interests are in computer architecture and its many affiliated areas, and he is co-leader of the TRIPS project at UT-Austin. |
| ||||
| ||||