- talks - lists.cs.princeton.edu

CITP Event Brad Smith – Immigration, Education, and the Future of Computer Science in America Friday, April 19, 12:30pm
by Nicole E. Wagenblast 17 Apr '13

17 Apr '13

Brad Smith – Immigration, Education, and the Future of Computer Science in America Date: Friday, April 19, 2013 Time: 12:30 PM Location: 3rd Floor Open Space, Sherrerd Hall Streaming Live: https://www.youtube.com/user/citpprinceton Brad Smith is Microsoft’s General Counsel and leads the company’s Department of Legal and Corporate Affairs (LCA). He is responsible for the company’s legal work, its intellectual property portfolio and patent licensing business, and its government affairs and philanthropic work. Since becoming general counsel in 2002, Smith has overseen numerous negotiations leading to competition law and intellectual property agreements with governments around the world and with companies across the IT sector. He has led the launch of global campaigns to bring enforcement actions against those engaged in software piracy and counterfeiting, malware, consumer fraud, and other digital crimes. Before joining Microsoft in 1993, Smith was a partner at Covington & Burling, having worked in the firm’s Washington, D.C., and London offices. He graduated summa cum laude from Princeton University and received his law degree at Columbia University.

1 0

Colloquium Speaker Ranjitha Kumar, Mon April 15, 4:30pm
by Nicole E. Wagenblast 12 Apr '13

12 Apr '13

Design Mining the Web Ranjitha Kumar , Stanford University Monday, April 15th, 4:30pm Computer Science 105 The Web has transformed the nature of creative work. For the first time, millions of people have a direct outlet for sharing their creations with the world. As a result, the Web has become the largest repository of design knowledge in human history, and the ensuing "democratization of design" has created a critical feedback loop, engendering a new culture of reuse and remixing. The means and methods designers employ to draw on prior work, however, remain mostly informal and ad hoc. How can content producers find relevant examples amongst hundreds of millions of possibilities and leverage existing design practice to inform and improve their creations? My research explores data-driven techniques for working with examples at scale during the design process, automating search and curation, enabling rapid retargeting, and learning generative probabilistic models to support new design interactions. Knowledge discovery and data mining have revolutionized informatics; in this talk, I'll discuss what we can learn from mining design. Ranjitha Kumar is a PhD candidate in the Computer Science Department at Stanford University, where she builds principled, data-driven tools for amplifying human creativity in design. Her work has received best paper awards or nominations at both of the premiere HCI conferences (CHI and UIST), and been recognized by the machine learning community through invited papers at IJCAI and ICML. She is the recipient of the 2011 Google PhD Fellowship in Design Development, and holds a BS in Computer Science from Stanford.

1 1

4:30pm Wed Apr 10 talk on NetFPGA platform in room CS 402
by Jennifer Rexford 10 Apr '13

10 Apr '13

Speaker: Andrew Moore (University of Cambridge) Title: NetFPGA: The Flexible Open-Source Networking Platform Time: 4:30pm Wednesday April 10 Place: CS 402 Abstract: The NetFPGA is an open platform enabling researchers and instructors to build high-speed, hardware-accelerated networking systems. The NetFPGA is the de-facto experimental platform for line-rate implementations of network research and it continues with a new generation platform capable of 4x10Gbps. The target audience is not restricted to hardware researchers: the NetFPGA provides the ideal platform for research across a wide range of networking topics from architecture to algorithms and from energy-efficient design to routing and forwarding. The most prominent NetFPGA success is OpenFlow, which in turn has reignited the Software Defined Networking movement. NetFPGA enabled OpenFlow by providing a widely available open-source development platform capable of line-rate and was, until its commercial uptake, the reference platform for OpenFlow. NetFPGA enables high-impact network research. This seminar will combine presentation and demonstration; no knowledge of hardware programming languages (eg Verilog/VHDL) is required. A NetFPGA 10G card will be awarded as a door-prize amongst the seminar attendees.

1 0

Colloquium Speaker Matei Zaharia Thurs April 11, 4:30pm
by Nicole E. Wagenblast 09 Apr '13

09 Apr '13

Making Big Data Analytics Interactive and Real-Time Matei Zaharia , University of California, Berkeley Thursday, April 11, 2013, 4:30pm Computer Science 105 The rapid growth in data volumes requires new computer systems that scale out across hundreds of machines. While early frameworks, such as MapReduce, handled large-scale batch processing, the demands on these systems have also grown. Users quickly needed to run (1) more interactive ad-hoc queries, (2) more complex multi-pass algorithms (e.g. machine learning and graph processing), and (3) real-time processing on large data streams. In this talk, we present a single abstraction, resilient distributed datasets (RDDs), that supports all of these emerging workloads by providing efficient and fault-tolerant in-memory data sharing. We have used RDDs to build a stack of computing systems including the Spark parallel engine, Shark SQL processor, and Spark Streaming engine. Spark and Shark can run machine learning algorithms and interactive queries up to 100x faster than Hadoop MapReduce, while Spark Streaming enables fault-tolerant stream processing at significantly higher scales than were possible before. These systems, along with several resource allocation and scheduling algorithms we have developed along the way, have been used in multiple industry and research applications, and have a growing open source community with 14 companies contributing in the past year. Matei Zaharia is a PhD student at UC Berkeley, working with Scott Shenker and Ion Stoica on topics in systems, cloud computing and networking. He is also a committer on Apache Mesos and Apache Hadoop. His work is supported by a Google PhD fellowship. Matei got his undergraduate degree at the University of Waterloo in Canada.

1 2

J Farnham preFPO
by Melissa M. Lawson 03 Apr '13

03 Apr '13

Jesse Farnham will present his preFPO on Friday April 12 at 9AM in Room 302 (note room!). The members of his committee are: Mona Singh, advisor; Olga Troyanskaya and Andrea LaPaugh, readers; Vivek Pai and Tom Funkhouser, nonreaders. Everyone is invited to attend his talk. His abstract follows below. --------------------------- "Organization and function acquisition in cellular networks" A main goal of biology is to understand the functional organization of organisms. In this thesis, we apply computational techniques to large, noisy biological datasets such as protein-protein interaction (PPI) networks and expression microarrays in order to elucidate principles of how genes in PPI networks acquire new functions and contribute to the overall functional organization of life. First, we investigate the functional organization of PPI networks by presenting a metric, functional insularity, that quantifies the degree to which a protein interacts with other proteins of similar function. We find high variation in insularity values, and we identify several differences between high- and low-insularity proteins. Secondly, we investigate function acquisition by human genes by identifying changes in protein-protein interactions, gene expression, and other markers of function as genes age. Our results indicate the presence of a tradeoff between modularity and connectivity in PPI network organization; furthermore, our results suggest that young genes acquire new functions gradually as they age, allowing the functional organization of the cell to evolve over time.

1 0

The Future of Internet Policy, Fri Apr 5 @ 4:30 in Sherrerd 101
by Stephen Schultze 02 Apr '13

02 Apr '13

Hello CS'ers, This is our very-occasional reminder that the Center for Information Technology Policy, across the quad in Sherrerd Hall, has many events of interest to you throughout the year. You can view and subscribe to our events calendar here: https://citp.princeton.edu/events/ This week, we have two events that you may want to check out: Thursday, April 4 at 12:30pm (lunch is provided) Grayson Barber -- Drones Are Like Flying Computers 306 Sherrerd Hall https://citp.princeton.edu/event/barber/ Friday, April 5 at 4:30pm (reception follows) Richard S. Whitt -- From SOPA to WCIT: A Glimpse of the Future of Internet Policy 101 Sherrerd Hall https://citp.princeton.edu/event/whitt/ - Richard Whitt is Corporate VP and Global Head of Public Policy and Government Relations, Motorola Mobility LLC, a subsidiary of Google Inc. Mr. Whitt will discuss how public policy grounded in the Internet’s architecture can best ensure that the Net fully enables tangible benefits such as innovation, economic growth, free expression, and user empowerment. In particular, recognizing that the Internet is rapidly becoming society’s chief operating system, Mr. Whitt will examine several recent examples of public policy threats to the Internet’s basic design principles. By contrast, an overarching public policy framework should be faithful to the multifaceted nature of the online world. Adhering to the deferential principle of “respect the functional integrity of the Internet,” in combination with the appropriate institutional and organizational implements, can help ensure that any potential regulation of Internet-based activities enables, rather than hinders, tangible and intangible benefits for end users. Rick oversees all interactions with government policymakers around the world at Motorola Mobility, a wholly-owned subsidiary of Google. Previously, he spent five and a half years in Google’s DC office, working on a variety of public policy issues. Most recently, he had served as director and managing counsel for federal policy, overseeing strategic thinking on privacy, cybersecurity, intellectual property, Internet governance, free expression, international trade, and telecom and media issues. Prior to joining Google in 2007, Rick founded and headed NetsEdge Consulting, a public policy consulting firm that provided legal analysis, regulatory strategy, and advocacy counsel. >From 1994 to 2006, Rick worked in the legal department at MCI Communications, where he served as vice president for federal law and policy. Rick previously spent over five years as an associate attorney in the communications practices of two large Washington, D.C.-based law firms. -- Stephen Schultze Associate Director, Center for Information Technology Policy at Princeton University https://citp.princeton.edu/

1 0

Colloquium Speaker Jacob Abernethy, Tues April 2, 4:30pm
by Nicole E. Wagenblast 02 Apr '13

02 Apr '13

Learning in an Adversarial World, with Connections to Pricing, Hedging, and Routing Jacob Duncan Abernethy , University of Pennsylvania Tuesday, April 2, 2013 , 4:30pm Computer Science 105 Machine Learning is often viewed through the lens of statistics, where one tries to model or fit a set of data under stochastic conditions; for example, it is typical to assume one's observations were sampled IID. However, dating back to results of Blackwell and Hannan from the 1950s we know how to construct learning and decision strategies that possess robust guarantees even under adversarial conditions. Within this setting the goal of the learner is to "minimize regret" against any sequence of inputs. In this talk we lay out the framework, discuss some recent results, and we finish by exploring a couple of surprising applications and connections, including: (a) market making in combinatorial prediction markets, (b) routing with limited feedback, and (c) hedging derivative securities (e.g. European option contracts) in the worst-case, with a connection to the classical Black-Scholes option-pricing model. Jake received his undergraduate degree in Mathematics from MIT in 2002 and a Master's degree in Computer Science from TTI-C in 2006. He recently finished a PhD in Computer Science at UC Berkeley, advised by Peter Bartlett, and he is now the Simons Postdoctoral Fellow at University of Pennsylvania. Jake has a particular focus on the intersection between machine learning, games and markets.

1 3

Colloquium speaker, Daniel Hsu, Thurs March 28, 4:30pm
by Nicole E. Wagenblast 28 Mar '13

28 Mar '13

Fast learning algorithms for discovering the hidden structure in data Daniel Hsu , Microsoft Research New England Thursday, March 28, 2013, 4:30pm Computer Science 105 A major challenge in machine learning is to reliably and automatically discover hidden structure in data with minimal human intervention. For instance, one may be interested in understanding the stratification of a population into subgroups, the thematic make-up of a collection of documents, or the dynamical process governing a complex time series. Many of the core statistical estimation problems for these applications are, in general, provably intractable for both computational and statistical reasons; and therefore progress is made by shifting the focus to realistic instances that rule out the intractable cases. In this talk, I'll describe a general computational approach for correctly estimating a wide class of statistical models, including Gaussian mixture models, Hidden Markov models, Latent Dirichlet Allocation, Probabilistic Context Free Grammars, and several more. The key idea is to exploit the structure of low-order correlations that is present in high-dimensional data. The scope of the new approach extends beyond the purview of previous algorithms; and it leads to both new theoretical guarantees for unsupervised machine learning, as well as fast and practical algorithms for large-scale data analysis. Daniel Hsu is a postdoc at Microsoft Research New England. Previously, he was a postdoc with the Department of Statistics at Rutgers University and the Department of Statistics at the University of Pennsylvania from 2010 to 2011, supervised by Tong Zhang and Sham M. Kakade. He received his Ph.D. in Computer Science in 2010 from UC San Diego, where he was advised by Sanjoy Dasgupta; and his B.S. in Computer Science and Engineering in 2004 from UC Berkeley. His research interests are in algorithmic statistics and machine learning.

1 3

Shi Li preFPO
by Melissa M. Lawson 25 Mar '13

25 Mar '13

Shi Li will present his preFPO on Thursday March 28 at 1:30 PM in Room 401 (note room!). The members of his committee are: Moses Charikar, advisor; Mark Braverman and Julia Chuzhoy (Toyota Tech Inst), readers; Zeev Dvir and Bernard Chazelle, nonreaders. Everyone is invited to attend his talk. His abstract follows below. ------------------------ title : Approximation Algorithms for Facility Location and Network Routing Problems abstract : My thesis is about approximation algorithms for two different classes of optimization problems. The first class is facility location problems. Two important problems in this category are uncapacitated facility location and k-median, both having long histories and numerous applications. We give improved approximation ratios for both problems. In the first part of my talk, I will focus on my recent work with Svensson (LS13) about an improved approximation algorithm for k-median. Our algorithm, which gives a 1+sqrt(3)+eps-approximation for k-median, is based on two rather surprising components. First, we show that it suffices to find an \alpha-approximate solution that contains k+O(1) medians. Second, we give such a pseudo-approximation algorithm with \alpha=1+sqrt(3)+eps. The second class is network routing problems. These are an important class of optimization problems, among which the Edge-Disjoint Paths (EDP) problem is one of the central and most extensively studied. In spite of its rich history, there is still a huge gap between the sqrt(log n)-hardness of approximation. In the second part of my talk, I will give an overview of my joint work with Chuzhoy (CL12), which gives a poly-logarithmic approximation for EDP by slightly relaxing the edge-disjointness constraint : we allow each edge in the network to be used twice (i.e, we allow congestion 2). This culminates a long line of research on the EDP with congestion problem.

1 0

Colloquium Speaker Daniel Licata Mon March 25, 4:30pm
by Nicole E. Wagenblast 25 Mar '13

25 Mar '13

Programming and Proving in Homotopy Type Theory Daniel Licata , Carnegie Mellon University & Institute for Advanced Study Monday, March 25, 2013, 4:30pm Computer Science 105 There is a strong correspondence between mathematics and programming, under which mathematical proofs correspond to programs. A _proof assistant_ is a tool that a mathematician/programmer can use to represent proofs/programs, in such a way that a computer can verify whether or not they are correct. The use of proof assistants in math and computer science is becoming ever more important for managing the increasing complexity of proofs and programs. Proof assistants have been used to check significant mathematical theorems, such as the Four Color Theorem and the Feit-Thompson Odd-Order Theorem, as well as large pieces of software, such as a C compiler and the definition of the Standard ML programming language. In this talk, I will describe my work on Homotopy Type Theory, which is a new foundation for proof assistants. Homotopy Type Theory extends type theory, the basis of several successful proof assistants, with new principles that bring it closer to informal mathematical practice. Moreover, it is based on an exciting new correspondence between type theory and the mathematical disciplines of homotopy theory and category theory, and consequently can be used to directly describe structures in these areas of math. I will describe computer-checked proofs of some basic results in algebraic topology, including calculations of homotopy groups of spheres and Eilenberg-MacLane spaces. Homotopy Type Theory is also a programming language, in which the new mathematical principles correspond to new programming techniques, which make certain programs easier to write. I will also describe my work on using Homotopy Type Theory as a programming language and its applications in computer science. Dan Licata received his PhD from Carnegie Mellon University in 2011, advised by Robert Harper. His dissertation, Dependently Typed Programming with Domain-Specific Logics, won the FoLLI E.W. Beth Dissertation Award in 2012. After graduating, he spent three semesters as a teaching fellow at Carnegie Mellon, designing and delivering a new introductory course on functional programming, parallelism, and verification. This year, he is a post-doctoral member at the Institute for Advanced Study, which is hosting a year-long program on an exciting new connection between programming languages and mathematics. His research focuses on creating better tools for computer-checked mathematics and software verification.

1 2