November 2012 - talks - lists.cs.princeton.edu

Colloquium Speaker Juan Bello Friday Nov 30, 3:30pm
by Nicole E. Wagenblast 29 Nov '12

29 Nov '12

Information Extraction from Music Audio Juan Bello , New York University Friday, November 30, 3:30pm Computer Science 105 This talk discusses a mix of concepts, problems and techniques at the crossroads of signal processing, machine learning and music. I will start by motivating the use of content-based methods for the analysis and retrieval of music. Then, I will introduce work in three projects being investigated in my lab: automatic chord recognition using hidden Markov models, music structure analysis using probabilistic latent component analysis, and feature learning using convolutional neural networks. In the process of doing so, I hope to illustrate some of the challenges and opportunities in the field of music informatics. Juan Pablo Bello is an Associate Professor of Music Technology at New York University. In 1998 he received his BEng in Electronics from the Universidad Simón Bolívar in Caracas, Venezuela, and in 2003 he earned a doctorate in Electronic Engineering at Queen Mary, University of London. Juan teaches and researches on the computer-based analysis of musical signals and its applications to music information retrieval, digital audio effects and interactive music systems. He is the principal investigator of the music informatics group of the Music and Audio Research Lab (MARL) at NYU. For more information please visit: https://files.nyu.edu/jb2843/public/ http://marl.smusic.nyu.edu/

1 3

Distinguished Colloquium Speaker Dan Spielman Thurs Nov 29 4:30pm
by Nicole E. Wagenblast 29 Nov '12

29 Nov '12

Algorithms, Graph Theory, and the Solution of Laplacian Linear Equations Dan Spielman , Yale University Thursday, November 29, 4:30pm Computer Science 105 We survey several fascinating concepts and algorithms in graph theory that arise in the design of fast algorithms for solving linear equations in the Laplacian matrices of graphs. We will begin by explaining why linear equations in these matrices are so interesting. The problem of solving linear equations in these matrices motivates a new notion of what it means for one graph to approximate another. This leads to the problem of approximating graphs by sparse graphs. Our algorithms for solving Laplacian linear equations will exploit surprisingly strong approximations of graphs by sparse graphs, and even by trees. We will survey the roles that spectral graph theory, random matrix theory, graph sparsification, low-stretch spanning trees and local clustering algorithms play in the design of fast algorithms for solving Laplacian linear equations. Daniel Alan Spielman received his B.A. in Mathematics and Computer Science from Yale in 1992, and his Ph.D in Applied Mathematics from M.I.T. in 1995. He spent a year as a NSF Mathematical Sciences Postdoc in the Computer Science Department at U.C. Berkeley, and then taught in the Applied Mathematics Department at M.I.T. until 2005. Since 2006, he has been a Professor at Yale University. He is presently the Henry Ford II Professor of Computer Science, Mathematics, and Applied Mathematics. He has received many awards, including the 1995 ACM Doctoral Dissertation Award, the 2002 IEEE Information Theory Paper Award, the 2008 Godel Prize, the 2009 Fulkerson Prize, the 2010 Nevanlinna Prize, an inaugural Simons Invesigator Award, and a MacArthur Fellowship. He is a Fellow of the Association for Computing Machinery and a member of the Connecticut Academy of Science and Engineering. His main research interests include the design and analysis of algorithms, graph theory, machine learning, error-correcting codes and combinatorial scientific computing.

1 2

Zach Ives visiting
by Ariel Rabkin 26 Nov '12

26 Nov '12

Hi All. Professor Zachary Ives is visiting us from the University of Pennsylvania. He will be giving a talk about his work in room 302 at 12:30 on December 5. The talk will be about ASPEN: Seamless Declarative Programming across Sensors, Streams, and the Cloud. Abstract: Due to advances in semiconductor and communications technology, we are steadily moving towards a world filled with a plethora of embedded sensing devices as well as "virtual" sensors within devices -- with the potential to monitor the environment and to react to or even anticipate changes. The environment being monitored might be a smart hospital, an entire campus, or even an office or data center. Often the behaviors we wish to produce involve combining constantly updating (streaming) data from multiple heterogeneous sensor and Internet sources. A major challenge is how to manage the resulting complexity, and to build large-scale, rich monitoring, learning, and control applications. In the ASPEN project we have been exploring an approach that seeks to separate logical dataflow from most of the algorithmic logic -- using a declarative, SQL-like (but iterative and incremental) programming model to capture the dataflow, data transformation, and state management needed by an application, combined with small bits of procedural code to handle complex logic. Our platform provides distributed query optimization that takes runtime conditions into account, while also supporting a range of learning, prediction, and connection-finding algorithms. In this talk I will describe our basic ASPEN prototype including cluster and sensor subsystems, and provide an overview of how we address issues of query optimization, distributed query execution, and incremental recomputation. Work done jointly with Mengmeng Liu, Svilen Mihaylov, Boon Thau Loo, and Sudipto Guha Speaker bio: Zachary Ives is an Associate Professor and the Markowitz Faculty Fellow at the University of Pennsylvania. His research interests include data integration and sharing, "big data", sensor networks, and data provenance and authoritativeness. He is a recipient of the NSF CAREER award, and an alumnus of the DARPA Computer Science Study Panel and Information Science and Technology advisory panel. He has also been awarded the Christian R. and Mary F. Lindback Foundation Award for Distinguished Teaching. He serves as the undergraduate curriculum chair for Penn's Singh Program in Market and Social Systems Engineering. He is a co-author of the textbook Principles of Data Integration.

1 0

4:30pm Mon Nov 26 talk on "Software-based services for cloud networks" in CS 402
by Jennifer Rexford 21 Nov '12

21 Nov '12

Software-based Services for Cloud Networks Anees Shaikh, IBM Research Monday, November 26, 2012, 4:30 PM - 5:30 PM Computer Science 402 Software-defined networking (SDN) aims to provide a well-defined programming and automation interface to network devices. In moving network control and functionality to software running on standard server platforms, SDNs depart from the traditional, vertically integrated model of network hardware and software. We view SDNs as a new opportunity for supporting more seamless integration of networks with IT processes, and for providing higher-value network services. In this talk, after describing the SDN model and our approach for building SDNs, we will focus on the problem of providing networking services for cloud computing platforms. As more enterprises look to leverage the cost and flexibility advantages of cloud computing, the lack of rich networking support remains a challenge. We will discuss the requirements of enterprise line-of-business applications for additional network functions in the cloud, and describe our research efforts to develop networking services for multi-tenant enterprise clouds using software-defined networking techniques. Anees Shaikh is a Research Staff Member and Manager with the IBM TJ Watson Research Center in New York. He currently leads the Systems Networking Research group at Watson, focusing on problems related to data center and cloud networks. www: http://researcher.ibm.com/researcher/view.php?person=us-aashaikh

1 0

Distinguished Colloquium Speaker Eva Tardos Monday Nov 19 4:30pm
by Nicole E. Wagenblast 19 Nov '12

19 Nov '12

How to Design Simple Efficient Mechanisms that are also Composable Eva Tardos , Cornell University Monday, November 19, 2012 , 4:30 PM - 5:30 P M Computer Science 105 E-commerce applications require simple, and well-designed systems, and systems that work well even if users participate in multiple mechanisms (and the value of each player overall is a complex function of their outcomes). Traditional mechanism design has considered such mechanisms only in isolation, and the mechanisms it proposes tend to be complex and impractical. In contrast, players typically participate in various mechanisms, mechanisms that are run by different principals (e.g. different sellers on eBay or different ad-exchange platforms) and coordinating them to run a single combined mechanism is infeasible or impractical. Even the simple and elegant Vickrey auction loses some of its appeal when not in isolation: when the overall value of each player is a complex function of their outcomes, the global mechanism is no longer truthful. In this talk we initiate the study of efficient mechanism design with guaranteed good properties even when players participate in multiple different mechanisms either simultaneously or sequentially. Over the last decade, computer scientists and game theorists have developed good understanding of the impact of strategic user behavior on system performance in environments that include selfish traffic routing, service location, and bandwidth sharing. We will consider simple mechanisms from this perspective. We define smooth mechanisms (that can be thought of as mechanisms that generate approximately market clearing prices), and show that smooth mechanisms are approximately efficient, and the efficiency guarantees for smooth mechanisms (when studied in isolation) carry over to the same or approximately the same guarantees for a market composed of such mechanisms. Joint work with Vasilis Syrgkanis. Eva Tardos is a Jacob Gould Schurman Professor of Computer Science at Cornell University, is Senior Associate Dean of Computing and Information Science, and was department chair 2006-2010. She received her PhD from Eotvos University in Budapest. She has been elected to the National Academy of Engineering and the American Academy of Arts and Sciences, and is the recipient of some fellowships and awards, including the Packard Fellowship, the Fulkerson Prize and the Dantzig prize. She was editor in chief of SIAM Journal of Computing 2003-09, and is editor of several other journals, including the Journal of the ACM, and Combinatorica. Her research interest is algorithms and algorithmic game theory, the subarea theory of designing systems and algorithms for selfish users.

1 2

Colloquium Speaker Ken Birman, Tues Nov. 13, 4:30pm
by Nicole E. Wagenblast 13 Nov '12

13 Nov '12

What are the Right Roles for Formal Methods in High Assurance Cloud Computing? Ken Birman , Cornell University Tuesday, November 13, 2012- 4:30pm Computer Science 105 The developer of a high assurance system must justify the fitness of the system for the intended use. In "small" scenarios, this problem is well understood: it entails rigorously specifying properties the system must achieve, designing protocols and correctness proofs, and then implementing and testing the needed code. In scaled-out settings, complex architectures are often unavoidable and this methodology can no longer be used. This talk focuses on how the resulting problems play out in the context of Isis2 (isis2.codeplex.com), a new system I've been building that supports a wide range of high-assurance properties for applications that will be hosted on cloud computing datacenters, and hence might need to run at very ambitious scale. We'll see that while it is possible to create systems of these kinds, existing formal methods are ill-suited to reasoning about the resulting solutions, suggesting that additional work will be needed to bridge between the mathematical and scientific foundations of high assurance of the kinds we teach in academic settings and the real-world instances of such problems that we may face in the near future. Ken Birman is the N. Rama Rao Professor of Computer Science at Cornell University, where he has headed a research effort in the area of high assurance distributed computing for thirty years. Ken is best known for inventing the virtual synchrony computing model and building the Isis Toolkit, which was ultimately used to build the system that operated the New York Stock Exchange for more than a decade, and the systems that continue to operate the French Air Traffic Control system and the US Navy AEGIS today. He also pioneered in the use of gossip protocols for system monitoring, management and control; several of his solutions are used today in the platforms that operate today's largest cloud computing infrastructures, notably at Amazon, IBM and Microsoft. A Fellow of the ACM since 1999, Ken won the 2009 IEEE Kanai Award for his research in distributed systems. He is currently developing online instructional materials for a short course on building high assurance cloud computing systems using his Isis2 platform.

1 2

Talk: Renata Teixeira, Broadband Internet Performance: A View From the Gateway, Mon Nov 12, 10am
by Nicole E. Wagenblast 12 Nov '12

12 Nov '12

Broadband Internet Performance: A View From the Gateway Renata Teixeira , University of Paris (LIP6) Monday, November 12, 2012, 10am Computer Science Room 402 Policymakers, ISPs, and users are increasingly interested in studying the performance of Internet access links. This talk will first present our findings on the the accuracy and overhead of state-of-the-art available bandwidth estimation tools to measure residential broadband throughput. Because of many confounding factors in a home network or on end hosts, however, thoroughly understanding access network performance requires deploying measurement infrastructure in users’ homes as gateway devices. This talk will then present the first study of network access link performance measured directly from home gateway devices. In conjunction with the Federal Communication Commission’s study of broadband Internet access in the United States, we study the throughput and latency of network access links using longitudinal measurements from nearly 4,000 gateway devices across 8 ISPs from a deployment of over 4,200 devices. We study the performance users achieve and how various factors ranging from the user’s choice of modem to the ISP’s traffic shaping policies can affect performance. Our study yields many important findings about the characteristics of existing access networks. Our findings also provide insights into the ways that access network performance should be measured and presented to users, which can help inform ongoing broader efforts to benchmark the performance of access networks. The work presented in this talk appeared at: -O. Goga, R. Teixeira, "Speed Measurements of Residential Internet Access", in Proc. of Passive and Active Measurement Conference, March 2012. -S. Sundaresan, W. de Donato, N. Feamster, R. Teixeira, S. Crawford, and A. Pescape, "Broadband Internet Performance: A View From the Gateway", in Proc. of ACM SIGCOMM, August 2011. Renata Teixeira received the B.Sc. degree in computer science and the M.Sc. degree in electrical engineering from Universidade Federal do Rio de Janeiro, Brazil, in 1997 and 1999, respectively, and the Ph.D. degree in computer science from the University of California, San Diego, in 2005. During her Ph.D. studies, she worked on Internet routing at the AT&T Research. She is currently a researcher with the Centre National de la Recherche Scientifique (CNRS) at LIP6, UPMC Sorbonne Universites, Paris, France. She was a visiting scholar at UC Berkeyley/ICSI. Her research interests are in measurement, analysis, and management of data networks. She has authored more than 40 papers in this area. Renata serves in the editorial board of the IEEE/ACM Transactions on Networking and of the ACM SIGCOMM Computer Communication Review. She is also a member of the steering committee of the ACM Internet Measurement Conference and has been active in the program committees of ACM SIGCOMM, ACM IMC, PAM, IEEE INFOCOM, among others.

2 1

Colloquium speaker Molly Steenson, Wednesday Nov, 7 at 4:30pm
by Nicole E. Wagenblast 07 Nov '12

07 Nov '12

"To the first machine that can appreciate the gesture:" Nicholas Negroponte and the MIT Architecture Machine Group Molly Steenson , Assistant Professor, School of Journalism & Mass Communication, University of Wisconsin-Madison PhD Wednesday, November 7, 2012 , 4:30pm Computer Science Small Auditorium (Room 105) Nicholas Negroponte today is best known for founding the MIT Media Lab in 1985. But before the Media Lab, he founded the Architecture Machine Group at MIT in 1968 (Arch Mach for short). As an architect captivated by computers, Negroponte created the lab as a test bed for interactive computers, sensors and programs that sought to change the manner in which computers and humans interacted with each other. He envisioned a world in which everybody would live in immersive, intelligent environments, even dedicating his first book The Architecture Machine “To the first machine that can appreciate the gesture." But what is the nature of that gesture? Arch Mach's structure and funding benefited from the same relationships as MIT’s major engineering and computing labs, collaborating especially closely with the Artificial Intelligence Lab, and its projects echoed the dynamics of Department of Defense funding that occurred in AI. In this talk, I will introduce Arch Mach's projects, projections—and gestures. Molly Wright Steenson is a Ph.D. candidate in Princeton University’s School of Architecture where she researches the history of architecture and information. In January 2013, she will join the faculty of the University of Wisconsin-Madison in the School of Journalism & Mass Communication, where she will focus on digital media studies. Steenson was also a professor at the Interaction Design Institute Ivrea in Italy. As a design researcher, she examines the effect of personal technology on its users, with recent projects in India and China for Microsoft Research and Intel Research and ReD Asscoiates. Prior to her scholarly career, Molly started working with the Web in 1994, producing sites and leading design strategy for Reuters, Netscape, and Razorfish. She holds a Master’s in Environmental Design from the Yale School of Architecture and a B.A. in German from the University of Wisconsin-Madison.

1 1

H Kim preFPO
by Melissa M. Lawson 06 Nov '12

06 Nov '12

Hanjun Kim will present his preFPO on Tuesday November 13 at 8:30 AM in Room 402. The members of his committee are: David August, advisor; David Wentzlaff and Kai Li, readers; JP Singh and Sharad Malik, nonreaders. Everyone is invited to attend his talk. His abstract follows below. ---------------------- ASAP: Automatic Speculative Acyclic Parallelization Speculative Decoupled Software Pipelining (Spec-DSWP) is a promising automatic parallelization technique that speculatively partitions a loop into multiple threads that communicate in a pipelined manner. Speculation can complement conservative static analysis, making automatic parallelization more robust and applicable. Pipelining allows Spec-DSWP to speculate only rarely occurring dependences while respecting the other dependences through communication among threads. Acyclic communication patterns in pipelining make the parallelized programs tolerant of communication latency. However, since Spec-DSWP partitions a loop iteration (a transaction) into multiple sub-transactions across multiple threads according to the pipeline stages, a special runtime system is required that supports multi-threaded transactions (MTX). This dissertation proposes Automatic Speculative Acyclic Parallelization (ASAP) that supports Spec-DSWP on real hardware without any hardware modiﬁcation. ASAP consists of a speculative acyclic parallelizing compiler and an MTX runtime system. The ASAP compiler automatically parallelizes loops using the Spec-DSWP transformation, and the ASAP runtime system correctly executes the speculatively transformed programs. With synergistic combination of speculation, acyclic communication, and runtime system support, ASAP achieves scalable performance for a wide range of applications on various parallel platforms such as multicore machines and clusters without any hardware modiﬁcation.

1 0