SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for November
-
Geppeto: Consumers Approach to Programming
Wed, Nov 05, 2008 @ 10:30 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Sinisa Srbljic, University of Zagreb, Croatia
Host: Prof. Nenad MedvidovicAbstract:
Contemporary society is experiencing a steady stream of new electronic gadgets, software products, and web applications. In this flood of functionality, users have adapted to rely less on manuals (if they are present at all) and shift their learning to trial and error, common paradigms, and experimentation. To accommodate this style of use . or perhaps driving this behavior - developers have successfully abstracted much of the technological complexity and transformed it into intuitive user interfaces often avoiding the need for reading lengthy manuals and formal training. Is it possible to adopt the same trial-and-error experimentation habit not only for using gadgets, but also for application development? We claim that intuitive aggregation and combination of software gadgets makes this possible.In this talk, we will show the use of current technology in building a consumer oriented development tool appropriate for individuals not formally trained in programming. We demonstrate that the complexity of existing system and scripting languages i.e.; syntax, semantics, control and data flow, data structures, data types, and programming components can be successfully replaced with analogies intuitively accessible to a much wider consumer population based exclusively on their use and understanding of user interfaces in popular web applications. We present a demo of Geppeto . a consumer tool for gadget-based application development. Composing gadgets with Geppeto does not require programming experience or reading of convoluted manuals. The presented research is sponsored by Google Inc. and the Croatian Ministry of Science.Biography:
Professor Sinisa Srbljic, Ph.D., is currently a professor at the School of Electrical Engineering and Computing, University of Zagreb, and the project leader of the Geppeto project. His career also spans Silicon Valley where he worked on large-scale distributed systems at AT&T Labs. He was visiting the University of Toronto, where he worked on the NUMAchine multiprocessor project, and the University of California, Irvine. His research interests include Web computing, gadget composition, and consumer programming. In teaching, he is involved in the theory of computing, programming language translation, service-oriented computing, and network middleware systems.Location: Waite Phillips Hall Of Education (WPH) - B27
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Computing Game-Theoretic Solutions
Thu, Nov 06, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Vincent Conitzer, Duke University
Host: Prof. Milind TambeAbstract:
Computer scientists are increasingly confronted with settings where multiple self-interested parties (humans or software agents) interact, especially in the context of the Internet. Examples include auctions, exchanges, elections, and other negotiation protocols, as well as job scheduling, routing, and webpage ranking. In these settings, the optimal course of action for one agent generally depends on what the other agents do, resulting in a tricky circularity. Game theory provides various notions of how agents should act in such domains. However, especially from an AI perspective, these concepts become useful only when we can compute the solutions that they prescribe. In this talk, I will review several standard game-theoretic solution concepts, including dominance, iterated dominance, Nash equilibrium, and Stackelberg strategies. I will also discuss algorithms and complexity results for computing these solutions.Biography:
Vincent Conitzer is an Assistant Professor of Computer Science and Economics at Duke University. He received Ph.D. (2006) and M.S. (2003) degrees in Computer Science from Carnegie Mellon University, and an A.B. (2001) degree in Applied Mathematics from Harvard University. His research focuses on computational aspects of microeconomics, in particular game theory, mechanism design, voting/social choice, and auctions. This work uses techniques from, and includes applications to, artificial intelligence and multiagent systems. Conitzer received an Alfred P. Sloan Research Fellowship (2008), an Honorable Mention for the 2007 ACM Doctoral Dissertation Award, the 2006 IFAAMAS Victor Lesser Distinguished Dissertation Award, the AAMAS Best Program Committee Member Award (2006), and an IBM Ph.D. Fellowship (2005). He is a co-author on papers that received a AAAI-08 Outstanding Paper Award and the AAMAS-08 Pragnesh Jay Modi Best Student Paper Award.Location: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
STAIR: The STanford Artificial Intelligence Robot project
Tue, Nov 11, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Andrew Y. Ng, Stanford University
Host: Prof. Fei ShaAbstract:
This talk will describe the STAIR home assistant robot project, and the satellite projects that led to key STAIR components such as (i) robotic grasping of previously unknown objects, (ii) depth perception from a single still image, and (iii) multi-modal robotic perception.Since its birth in 1956, the AI dream has been to build systems that exhibit broad-spectrum competence and intelligence. STAIR revisits this dream, and seeks to integrate onto a single robot platform tools drawn from all areas of AI including learning, vision, navigation, manipulation, planning, and speech/NLP. This is in distinct contrast to, and also represents an attempt to reverse, the 30 year old trend of working on fragmented AI sub-fields. STAIR's goal is a useful home assistant robot, and over the long term, we envision a single robot that can perform tasks such as tidying up a room, using a dishwasher, fetching and delivering items, and preparing meals.In this talk, I'll describe our progress on having the STAIR robot fetch items from around the office, and on having STAIR take inventory of office items. Specifically, I'll describe: (i) learning to grasp previously unseen objects (including unloading items from a dishwasher); (ii) probabilistic multi-resolution maps, which enable the robot to open/use doors; (iii) a robotic foveal+peripheral vision system for object recognition and tracking. I'll also outline some of the main technical ideas---such as learning 3-d reconstructions from a single still image, and reinforcement learning algorithms for robotic control---that played key roles in enabling these STAIR components.Biography:
Andrew Ng is an Assistant Professor of Computer Science at Stanford University. His research interests include machine learning, reinforcement learning/control, and broad-competence AI. His group has won best paper/best student paper awards at ACL, CEAS, 3DRR and ICML. He is also a recipient of the Alfred P. Sloan Fellowship.Location: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Robot Babies
Thu, Nov 13, 2008 @ 12:30 PM - 01:50 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Melinda Snodgrass
Host: Prof. Milind Tambe
This lecture is part of CS499 "Intelligent agents and science fiction". Abstract:
This talk is designed to raise several questions.The first is to examine the purpose of the robot in fiction. Aliens can also serve this purpose because they, like robots, are The Other. Ultimately we use the robot as a way to explore issues that confront humans. What does it mean to be human? What is morality?In fiction we often treat the robot as a blank slate or even as a child learning the ways of the world. Which leads to the second question --Why build an AI? Is this driven by the human need to create? A night of unprotected sex can accomplish that. Why build a machine? I would like to involve the audience in this part of the discussion.Finally I will look at how I find dramatic stories regarding robots/AI.s, and also how the stories have changed as our world has become ever more wired. You notice there aren.t very many stories about power hungry computers taking over the world any longer because we.ve all faced the Blue Screen of Death. We live and work with computers every day, and we know they are fundamentally stupid.So, if we build an AI do we make it smart enough to be venal?Biography:
Melinda Snodgrass was born in Los Angeles, but her family moved to New Mexico when she was five months old so she considers herself a native. She was educated at the University of New Mexico graduating Magna cum Lauda in History. During her undergraduate days she took a year off to study opera at the Conservatory of Vienna in Vienna, Austria. Upon her return from Europe she entered the U.N.M. School of Law.After graduation she practiced for three years, and discovered that while she loved the study of law she didn't particularly love lawyers so she began a career in writing. She has written numerous science fiction novels, and helps edit and writes for the WILD CARD anthologies.In 1988 she accepted a job on Star Trek: The Next Generation, and began her Hollywood career. Her most recent Hollywood position was as Consulting Producer on the N.B.C. show PROFILER. She has written television pilots and feature films.Her novel, THE EDGE OF REASON, an occult thriller, has just been released by Tor books, and the second novel THE EDGE OF RUIN will be published in June of 2009. she is hard at work on her story for the next Wild Card's Anthology, and she is working with Ian Tregillis to adapt his novel, BITTER SEED, as a screenplay. Her passion (aside from writing) is riding Grand Prix dressage on her Lusitano stallion Vento da Broga.Location: Seeley Wintersmith Mudd Memorial Hall (of Philosophy) (MHP) - 101
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Optimizing Locomotion: Learning Control at Intermediate Reynolds Number
Fri, Nov 14, 2008 @ 02:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Russ Tedrake, MIT
Host: Prof. Gaurav SukhatmeAbstract:
In this talk I'll describe our efforts in using computational tools from optimal control theory (including machine learning and motion planning algorithms) to design efficient and agile control systems for locomotion. In particular, I'll emphasize new results applying these ideas to bird-scale aerial vehicles in complicated fluid regimes. Fluid dynamics, particularly at intermediate Reynolds numbers, represents one of the hard sciences where experiments still have a clear advantage over theory. It also happens that this dynamic regime offers some incredibly exciting controls problems; problems where classical control approaches have not made significant progress. I will argue that optimal control methods based on approximate models and model-free reinforcement learning methods are very well-suited to these regimes and may be the most natural route to finding efficient, high-performance control solutions. I'll describe our learning experiments with robotic birds (which fly with flapping wings) and with an airplane that can land on a perch. I will also briefly describe how these tools can be applied naturally to the control of minimally-actuated walking machines on rough terrain.Biography:
Russ Tedrake is the X Consortium Assistant Professor of Electrical Engineering and Computer Science at MIT, and a member of the Computer Science and Artificial Intelligence Lab. He received his B.S.E. in Computer Engineering from the University of Michigan, Ann Arbor, in 1999, and his Ph.D. in Electrical Engineering and Computer Science from MIT in 2004, working with Sebastian Seung. After graduation, he joined the MIT Brain and Cognitive Sciences Department as a Postdoctoral Associate. In 2008, he received an NSF CAREER award, the MIT Jerome Saltzer award for undergraduate teaching, and was named a Microsoft Research New Faculty Fellow.Location: Ronald Tutor Hall of Engineering (RTH) - 406
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Improving deep packet inspection through extended automata
Tue, Nov 18, 2008 @ 11:00 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Cristian Estan, University of Wisconsin Host: Prof. Ramesh Govindan Abstract:
Deep packet inspection is playing an increasingly important role in novel network services. Regular expressions are the language of choice for writing signatures used in deep packet inspection, but standard signature matching solutions are not suitable for high-speed environments. Deterministic finite automata (DFAs) are fast but combining the DFAs for multiple signatures often leads to state space explosion. Non-deterministic finite automata (NFAs) are small but matching can be slow for large signature sets. This talk presents a new solution that simultaneously addresses these problems. Extended finite automata (XFAs) augment deterministic finite automata (DFAs) with finite auxiliary variables and simple instructions that manipulate them. The introduction of auxiliary variables allows us to eliminate state space explosion. In experiments with signature sets used for intrusion prevention by Snort and Cisco Systems, XFAs simultaneously reduce memory and run time by more than an order of magnitude when compared to earlier solutions. Biography:
Cristian Estan has been an assistant professor in the Computer Sciences Department at University of Wisconsin-Madison since Fall 2004. His Ph.D. is from University of California, San Diego (adviser George Varghese). His research focuses on network security, network traffic measurement, and network traffic analysis. It has resulted in publications in top conferences in networking, security, systems, and databases: SIGCOMM, IEEE Security and Privacy (Oakland), OSDI, SIGMETRICS, ICDE, IMC, etc. His work is supported by multiple grants from NSF including a CAREER grant and gifts from Cisco Systems.Location: Charles Lee Powell Hall (PHE) - 223
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
On Broadcast Scheduling
Tue, Nov 18, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Samir Khuller, University of Maryland
Host: Prof. Leana GolubchikAbstract:
Broadcasting over a wireless channel is a standard mechanism to disseminate information to a set of clients. Clients request different pieces of information, called "pages", and in this "pull-based" model, the server responds with a broadcast schedule. The key property that distinguishes this problem from standard scheduling is that multiple requests may be satisfied by a single broadcast of the requested page. Surprisingly, this small change makes almost all the problems in this area NP-hard whereas without this property these problems can be solved optimally in polynomial time for unit sized pages. This overlap property arises in other contexts as well such as in multi-query processing.We consider a variety of different objective functions in our work minimizing the sum of response times, minimizing the maximum response time and maximizing the number of satisfied requests when requests have deadlines. This is a survey talk based on several papers and will cover a collection of results using a variety of techniques. Finally we propose some open problems. No background knowledge beyond undergraduate algorithms is expected.Biography:
Samir Khuller received his M.S and Ph.D from Cornell University in 1989 and 1990, respectively. He spent two years as a Research Associate at the Institute for Advanced Computer Studies (UMIACS) at the University of Maryland, before joining the Computer Science Department in 1992, where he is currently a Professor in the Department of Computer Science. He spent several summers at the IBM T. J. Watson Research Center, and also visited the IBM Tokyo Research Lab for several weeks. From 2004 to 2008 he was the Associate Chair for Graduate Education.His research interests are in graph algorithms, discrete optimization, and computational geometry. He has published about 140 journal and conference papers, and several book chapters on these topics. He is an editor for the journal Networks, International Journal on Foundations of Computer Science, problems Editor for ACM Trans. on Algorithms, and a columnist for SIGACT News. He has served on several program committees.He received the National Science Foundation's Career Development Award, several Dept. Teaching Awards, the Dean's Teaching Excellence Award and also a CTE-Lilly Teaching Fellowship. In 2003, he and his students were awarded the "Best newcomer paper" award for the ACM PODS Conference. He received the University of Maryland's Distinguished Scholar Teacher Award in 2007, as well as a Google Research Award. He graduated at the top of the Computer Science Class from IIT-Kanpur.Location: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Stochastic Local Search for Propositional Satisfiability
Mon, Nov 24, 2008 @ 02:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Prof. Abdul Sattar, Griffith University, Australia
Host: Prof. Milind TambeAbstract:
The problem of finding a consistent truth assignment to all propositional variables in a formula, known as SAT problem, has been an interesting and difficult challenge. Indeed, SAT is at the heart of all computationally intractable problems. Many real world problems could be encoded as SAT problems. Thus finding an efficient solution for SAT has far reaching impact on computationally hard problems. This talk will begin with an overview of the main approaches for solving SAT problems. We will then focus on stochastic local search based methods. These methods have been shown to be highly effective for large size problems. We will present our recent results on clause weighting based local search, including an influential method that automatically learns about the structure of the problem, and efficiently exploit those structures to solve some of the difficult challenge problems in the field.Biography:
Prof Sattar is founding Director of the Institute for Integrated and Intelligent Systems (IIIS), a research centre of excellence at Griffith University established in 2003. He is a Research Leader at NICTA Queensland Laboratory since June 2005, and also held the Associate Director of Education portfolio at the Queensland Laboratory from October 2006-June 2008. He has been an academic staff member at Griffith University since February 1992 as a lecturer (1992-95), senior lecturer (1996-99), and professor (2000-present) within the School of Information and Communication Technology. Prior to his career at Griffith University, he was a lecturer in Physics in Rajasthan, India (1980-82), research scholar at Jawaharlal Nehru University, India (1982-85), the University of Waterloo, Canada (1985-87), and the University of Alberta, Canada (1987-1991).He holds a BSc (Physics, Chemistry and Mathematics) and an MSc (Physics) from the University of Rajasthan, India, an MPhil in Computer and Systems Sciences from the Jawaharlal Nehru University, India, and an MMath in Computer Science from the University of Waterloo, Canada, and a PhD in Computer Science (with specialization in Artificial Intelligence) from the University of Alberta, Canada. His current research interests include knowledge representation and reasoning, constraint satisfaction, rational agents, propositional satisfiability, temporal reasoning, temporal databases, and bioinformatics. He has supervised 17 successful completions of PhD graduates, and published over 100 technical papers in refereed conferences and journals in the field. His research team has won three major international awards in recent years (the gold medals for the SAT 2005 and SAT 2007 competitions in the random satisfiable category and an IJCAI 2007 distinguished paper award).Location: Kaprielian Hall (KAP) - 144
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
The Promise, the Limits, and the Beauty of Software (Distinguished Lecture)
Mon, Nov 24, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dr. Grady Booch, IBM Research
Host: Prof. Barry BoehmAbstract:
Within this generation, software has changed the way that individuals collaborate, organizations do business, economies operate, and cultures interact. Software-intensive systems can amplify human intelligence, but they cannot replace human judgment; software-intensive systems can fuse, coordinate, classify, and analyze information, but they cannot create knowledge.Although software offers seemingly limitless promise, there are some very real limits to what software can do. Not everything we want to build can be built: there exist pragmatic theoretical and technical limits that make software development hard, if not in some cases impossible.Furthermore, not everything we want to build should be built: there exist moral, economic, social, and political limits that govern human industry.Software-intensive systems are perhaps the most intellectually complex artifacts created by humans, and while the majority of individuals in the civilized world rely on software in their daily lives, few of them understand the essential complexity therein, the labour required to create such artifacts, and the beautiful and elegant chaos of their architecture.In this presentation, we will examine the promise, the limits, and the beauty of software, as well as offer some conclusions that can be drawn from the last 60 years of software and some expectations and cautions for the next generation.Biography:
http://www.handbookofsoftwarearchitecture.com/index.jsp?page=ContactLocation: Olin Hall of Engineering (OHE) - 132
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Forget him (or her) and keep on moving: Making mobile social networks navigable
Tue, Nov 25, 2008 @ 11:00 AM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Augustin Chaintreau, Thomson, Paris
Host: Prof. Ramesh GovindanAbstract:
A network is navigable if a simple decentralized scheme allows efficient routing (i.e. in a polylogarithmic number of steps). Previous works have shown that general classes of graphs can be made navigable by adding few links according to an appropriate distribution. However, for most of this graphs, navigability is sensitive to small deviations from this distribution. Moreover, it seems difficult for the nodes to manage such a link addition in a distributed way. In spite of some efforts, and evidence of the "small-world phenomenon" in social networks, no model currently proves the emergence of navigability from local dynamics.Here we prove that navigability emerges from nodes own mobility and memory. Inspired by emerging opportunistic mobile networks using human carried devices (a.k.a. Pocket switched networks), we model a network where nodes move (in our case, according to a random walk in dimension d), and may opportunistically create connections as they meet physically. Once established, these connections are randomly maintained or forgotten, based only their current age. We prove that this simple setting allows one to create navigable networks. We present a few applications of these techniques to design opportunistic spatial gossip, and discuss the upcoming challenges in relation with recent experiences on using social software for opportunistic mobile networks.(this is a joint work with Pierre Fraigniaud and Emmanuelle Lebhar, from CNRS-Universite Paris Diderot and CMM-Universidad de Chile)Location: Kaprielian Hall (KAP) - 163
Audiences: Everyone Is Invited
Contact: CS Colloquia
