February 2007 - talks - lists.cs.princeton.edu

Multi-core processing lecture at 1:30 today
by Jaswinder Pal Singh 27 Feb '07

27 Feb '07

Prof Kunle Olukotun from Stanford is giving a guest lecture on multi-core processor architecture and software/models in COS 598A today, from 1:30 to 3 PM. Room coming shortly. This may be of interest to a broader audience given the strong multicore trend in processors going forward, and Prof. Olukotun is a leading researcher in the area. All are welcome to attend. Cheers, JP

1 1

E Nabieva preFPO
by Melissa M Lawson 23 Feb '07

23 Feb '07

To correct my earlier announcement, Elena's preFPO is Tuesday (not Thursday) February 27. All other details remain the same. --Melissa

1 0

E Nabieva preFPO
by Melissa M Lawson 22 Feb '07

22 Feb '07

Elena Nabieva will present her preFPO on Thursday February 27 at 11:15AM in Room 253, Carl Icahn Lab. The members of her committee are: Mona Singh, advisor; Tom Funkhouser, Olga Troyanskaya, readers; Bernard Chazelle, Ned Wingreen (MolBio), nonreaders. Everyone is invited to attend her talk. Her abstract follows below. ------------------------------ Abstract: Exploring the interplay between topology and function in protein interaction networks The emergence in recent years of numerous high-througphut experimental techniques in biology has lead to a new, genome-scale approach towards biological research. This high-throughput biology faces two complementary tasks: obtaining data on genomic scale and making sense of this data. It is the second task where computer scientists working in computational biology can make great contribution. One type of data obtained by high-throughput experiments is information about interactions among proteins, such as physical protein-protein interactions. This information can bring scientists closer to a solution to one of the most important problems in biology: understanding the role that different proteins play in the cell and the interplay among them. In my work, I look at the relationship between protein function and the protein's context in the interaction network from two angles: using interaction networks and information about other proteins to predict a protein's cellular role, and finding schemas, or recurring patterns of interaction among different types of proteins. In the first part of the talk, I explore the use of physical protein interaction networks for predicting the function of proteins. First, using as illustration some of the existing approaches to this problem, I discuss which topological properties of interaction networks should be taken into account by algorithms for predicting protein function based on physical interaction networks. Using these desiderata as guidelines, I introduce an original network-flow based algorithm called FunctionalFlow that exploits the underlying structure of protein interaction maps in order to predict protein function. In cross-validation testing on the yeast proteome, I show that FunctionalFlow has improved performance over previous methods in predicting the function of proteins with few (or no) annotated protein neighbors. I demonstrate that FunctionalFlow performs well because it takes advantage of both network topology and some measure of locality. Finally, I show that performance can be improved substantially as we consider multiple data sources and use them to create weighted interaction networks. In the second part of the talk, I take a different view at the topology-function relationship and use known information about protein molecular function and the physical interaction network to attempt to uncover organizational principles of the network. In this bottom-up view, I examine the networks from the perspective of ``pathway schemas,'' or recurring patterns of interaction among different types of proteins. Proteins in these schemas tend to act as functional units within diverse biological processes. I discuss computational methods for automatically uncovering statistically over-represented pathway schemas in protein-protein interaction maps, and touch upon the comparative-interactomics aspects of this problem. Coming back to the task of improving our understanding of protein function, I conclude by demonstrating how overrepresented schemas can be used to gain new insights about the biological function of proteins.

1 0

Speaker Tomorrow
by Ginny Hogan 20 Feb '07

20 Feb '07

Unfortunatley, David Forsyth are colloquium speaker for tomorrow had to cancel his trip - there will be no colloquium talk - Ginny

1 0

Colloquium Speaker 2/21/2007
by Ginny Hogan 19 Feb '07

19 Feb '07

David Forsyth, UIUC, will present his Talk, "Looking at People" Wed, Feb. 21st at 4:15pm COS 105 -

1 0

Today: Computational Functional Genomics or "What to do with all this Data?"
by Chrissy Colvin 14 Feb '07

14 Feb '07

** PICASso: ** Program in Integrative Information, Computer and Application Sciences ** www.cs.princeton.edu/picasso ** ** TALK TODAY--Wednesday, February 14, 2007 Computation and Data Analysis in Biology and Information Sciences www.cs.princeton.edu/picasso/bio_info_lunch.php TITLE: Computational Functional Genomics or "What to do with all this Data?" SPEAKER: Olga Troyanskaya, Computer Science & Lewis Sigler Institute, Princeton University TIME: Wednesday, February 14, 2007 Seminar begins at 12:30 p.m. (lunch provided ~12:20) LOCATION: Room 402, Computer Science, Princeton University ABSTRACT: Assigning protein function and modeling their interactions in biological networks is a key challenge in modern systems biology. Broad availability of diverse functional genomic data such as microarray and protein interaction studies should enable fast and accurate generation of functional predictions and network models through computational prediction and experimental validation. However, these data remain vastly underutilized due to their biological and experimental diversity, high noise levels, and computational and semantic complexity. I will discuss challenges and offer some computational solutions in extracting from these functional genomic data biological knowledge about protein function, regulation, and interactions. SIGN UP FOR THE PICASSO MAILING LIST: ===================================== If you would like to be kept informed of computationally-oriented events in (and around) Princeton, please SUBSCRIBE to the PICASso mailing list by visiting https://lists.cs.princeton.edu/mailman/listinfo/picasso. This page also contains information on how to UNSUBSCRIBE. PLEASE FORWARD THIS MESSAGE TO OTHER COMPUTATIONALLY-ORIENTED RESEARCHERS WHO MAY BE INTERESTED IN THIS EVENT, OR FUTURE PROGRAMS. THANKS!

1 0

Ge Wang preFPO
by Melissa M Lawson 12 Feb '07

12 Feb '07

Ge Wang will present his preFPO on Friday February 16 at 2:00pm in Room 402. The members of his committee are: Perry Cook, advisor; Paul Lansky (Music) and Roger Dannenberg (CMU), readers; Andrew Appel and Brian Kernighan, non-readers. Everyone is invited to attend his talk. His abstract follows below. ------------------------------------- The ChucK Audio Programming Language ("A Strongly-timed and On-the-fly Environ/mentality") The computer has long been considered an extremely attractive tool for creating, manipulating, and analyzing sound. Its precision, possibilities for new timbres, and potential for fantastical automation make it a compelling platform for expression and experimentation - but only to the extent that we are able to express to the computer what to do, and how to do it. To this end, the programming language has served, perhaps, as the most general, and yet most precise and intimate interface between humans and computers. Furthermore, "domain-specific" languages can bring additional expressiveness, conciseness, and perhaps even different ways of thinking to the domain (e.g. sound synthesis). This thesis argues for the philosophy, design, and development of ChucK, a domain-specific programming language designed for computer music. The goal is to create a language that is "expressive" and "easy to write and read" with respect to time and parallelism, and to provide a platform for precise synthesis/analysis and rapid experimentation in computer audio. In particular, ChucK provides a syntax for representing information flow, a new time-based concurrent programming model that allows programmers to easily (and flexibly) control the flow of time in code (we call this "strongly-timed"), and facilities to develop programs "on-the-fly" - as they run. A ChucKian approach to "live coding" as a new musical performance paradigm is also described - and ultimately motivates the Audicle, a specialized graphical environment designed to facilitate on-the-fly programming, to visualize and monitor ChucK programs in real-time, and to provide a platform for building highly customizable user interfaces. The various aspects of the ChucK language (and Audicle) are evaluated in the context of computer music research, performance, and pedagogy. As part of an extensive case study, the thesis discusses ChucK as a primary teaching and development tool in the Princeton Laptop Orchestra (PLOrk), which continues to be a wonderfully useful platform to observe the effectiveness of ChucK for teaching topics ranging from programming to sound synthesis, and for crafting new instruments, compositions, and performances for computer-mediated ensembles. Additional applications are also described, including classrooms, live coding "arenas", compositions and performances, user studies, and integrations of ChucK into other software systems, such as the sound design environment TAPESTREA. The oral presentation shall also include a brief history of computer music and programming, and offer contrast between ChucK and other computer music languages and environments.

1 0

S Gaw preFPO
by Melissa M Lawson 09 Feb '07

09 Feb '07

Shirley Gaw will present her preFPO on Thursday February 15 at 4:30PM in Room 302 (note room). The members of her committee are Ed Felten, advisor; Paul Dourish (UC Irvine) and Brian Kernighan, readers; Perry Cook and Olga Troyanskaya, non-readers. Everyone is invited to attend her talk. Her abstract follows below. ------------------------------------- Human factors are considered one of the "weakest links" in computer security. That is, there is a gap between the theoretical security of a system and the practical security of a system. My thesis addresses this gap by understanding how users observe or avoid secure practices. The central argument of my thesis is that a sociological approach to computer security is a missing but necessary foundation for addressing the "weakest link" in secure systems. While we may have intuitions about human behavior, simplistic models of users fail to capture the underlying complexity of their practices in real situations, including both their adoption of security technologies and their implementation of security policies. I will illustrate this by discussing results from three studies investigating the human side of computer security: use of encrypted e-mail in an activist organization, password management practices of undergraduates, and privacy concerns of knowledge workers. Finally, I will discuss how results from sociological approaches can inspire design for novel systems. We are working on using the results from the encrypted e-mail study to design a communication system where interaction relies on implicit cues of user involvement and trust built over time.

1 0

Optimization Techniques in Linux Development Feb 20-22
by Chrissy Colvin 07 Feb '07

07 Feb '07

Registration is required and space is limited. To register, email ccolvin(a)cs.princeton.edu DATES: February 20, Friend 004 February 21, Friend 009 February 22, Friend 004 TIME: 4:30 pm - 7:00 PM Linux Performance measurement and optimization is a black art: to achieve results, knowledge of a varied set of tools and insight into the system's internals are required to successfully conclude the investigation with clear, irrefutable evidence of the responsible bottleneck. The speaker will take the audience through a marathon demonstration of the primary benchmarking tools, highlighting their features, pitfalls and limits, while at the same time describing the underlying system mechanisms interrelating with what is being measured. Initially focusing on the performance of the whole system, the discussion will then shift to performance measurement and optimization of a specific process, covering the approaches available to benchmark and tune programs running natively on the system or inside a virtual machine. Three sessions will cover the necessary theory and introductory hands-on practice, as we cover both System and Network performance problems. Instructor: Federico Lucifredi, Harvard University Federico Lucifredi is the maintainer of the man suite, the primary documentation-delivery tool under Linux, a graduate student at Harvard University, and a software-engineer-turned-manager at Novell corporation. Previously, Federico has been a CTO and a network software architect '.com' and embedded Linux startups, and he has spent two years teaching in Boston University's graduate and undergraduate programs, while simultaneously consulting for MIT. He is a frequent speaker at user group and conference events, notably the LinuxWorld and the IMPlanet conferences, where he was a panelist representing the jabber community. Federico is a recognized expert in computing performance issues, and consults pro-bono with Standard and Poor's clients interested in Free/Open Source technical and strategic issues, while currently participating in the GPL v3 drafting process in the large-corporation panel. Registration is required and space is limited. To register, email ccolvin(a)cs.princeton.edu DATES: February 20, Friend 004 February 21, Friend 009 February 22, Friend 004 TIME: 4:30 pm - 7:00 PM

1 0

M Dudik preFPO
by Melissa M Lawson 07 Feb '07

07 Feb '07

Miro Dudik will present his preFPO on Friday February 9 at 10AM in Room 402. The members of his committee are Rob Schapire, advisor; David Blei and Stephen Phillips (AT&T), readers; Moses Charikar and Olga Troyanskaya, nonreaders. Everyone is invited to attend his talk. His abstract follows below. ----------------------------------------- Maximum entropy, generalized regularization, and modeling species habitats Maximum entropy (maxent) approach, equivalent to maximum likelihood, is a widely used method for estimating probability distributions. However, when trained on small datasets, maxent is likely to overfit. Therefore, many smoothing techniques were proposed to mitigate overfitting. In my dissertation, I propose a unified treatment for a large and general class of smoothing techniques including L1 and L2 regularization. As a result, it is possible to prove non-asymptotic performance guarantees and derive novel regularizations based on structure of the sample space. To obtain solutions for a large class of maxent problems, I propose new algorithms derived from boosting and iterative scaling. Convergence of these algorithms is proved using a novel method, which unifies previous approaches based on information geometry and compactness. As an application of maxent, I discuss an important problem in ecology: modeling distributions of biological species. Regularized maxent fits this problem well and offers several advantages over previous techniques. In particular, it addresses the problem in a statistically sound manner and allows principled extensions to situations when data is collected in a biased manner or when we have access to data on many related species. The utility of maxent is demonstrated on large real-world datasets.

1 0