March 2010 - talks - lists.cs.princeton.edu

Colloquium Speaker, Tues April 6- Rebecca Fiebrink
by Nicole E. Wagenblast 02 Apr '10

02 Apr '10

Real-time Human-Computer Interaction with Supervised Learning Algorithms for Music Composition and Performance Rebecca Fiebrink, Princeton University, Computer Science Department Tuesday, April 6, 2010- 4:30pm Small Auditorium CS105 Supervised learning offers a useful set of algorithmic tools for many problems in computer music composition and performance. Through the use of training examples, these algorithms offer human musicians a means to implicitly specify the relationship between low-level, human-generated control signals (such as gesturally-manipulated sensor outputs or audio captured by a microphone) and the desired computer response (such as a change in synthesis or structural parameters of dynamically-generated audio). In my work, I explore how to most effectively enable users to interact with supervised learning algorithms to compose and perform new music. I have built a general-purpose software system for applying standard supervised learning algorithms in real-time problem domains. This system, called the Wekinator, supports human interaction throughout the entire supervised learning process, including the generation of training examples and the application of trained models to real-time inputs. Already, the Wekinator has enabled the creation of several new compositions and instruments. Furthermore, this system has enabled me to study several aspects of human-computer interaction with supervised learning in computer music. I have used the Wekinator as a foundation for a participatory design process with practicing composers, ongoing work with non-expert users in a classroom context, and the design of a gesture recognition system for a sensor-augmented cello bow. This research has led to a clearer characterization of the requirements and goals of instrument builders and composers, a better understanding of how to design user interfaces for supervised learning in both real-time and creative application domains, and a greater insight into the roles that interaction (encompassing both human-computer control and computer-human feedback) can play in the development of systems containing supervised learning components. This work highlights how music and other creative endeavors differ from more traditional applications of supervised learning, and it contributes to a broader HCI perspective on machine learning practice.

1 1

Korhammer Lecture: The Art of Green Design: Doing Nothing Well - Prof. Jan M. Rabaey, UC Berkeley
by Jennifer Rexford 25 Mar '10

25 Mar '10

> > > KORHAMMER LECTURE > > PRINCETON UNIVERSITY, DEPARTMENT OF ELECTRICAL ENGINEERING > > “The Art of Green Design: Doing Nothing Well” > > Jan M. Rabaey, Donald O. Pederson Distinguished Professor, University of California, Berkeley > > Monday, March 29, 2010, 4:30pm, Computer Science 104 > > Abstract: > Reducing human appetite for energy and the corresponding carbon footprint is one of the dominant societal problems of today. Information technology (IT) very likely will play a major role in our quest to generate, distribute, store and consume energy more efficiently. Unfortunately the IT platform itself represents a major consumer of electrical power, and its energy footprint is very likely to continue to grow over the next decades. > > The most obvious approach to address this challenge is to continue to improve the energy efficiency of our computation, communication or storage devices. Unfortunately, the opportunities of doing so (using technology scaling or innovative architectures) are getting increasingly limited. Instead, a lower-hanging fruit is to eliminate the massive amount of waste that is inherent to today’s IT systems. Systems tend to be designed and optimized for peak performance. In reality, most computation nodes, networks and storage devices typically operate at a fraction of the maximum load, and do this with surprisingly low energy efficiency. If we could design systems that do “nothing” well (as phrased by David Culler), major energy savings would be enabled. Accomplishing “energy-proportional” computing requires a full-fledged top-down and bottom-up approach to the design of IT systems. This talk discusses what this means for a couple of systems ranging from large scale data centers over communication networks to small-scale body-area networks for medical applications. > > > > >

1 0

Colloquium Speaker- Thurs. March 25- Vinod Vaikuntanathan
by Nicole E. Wagenblast 23 Mar '10

23 Mar '10

Side Channels and Clouds: New Challenges in Cryptography Vinod Vaikuntanathan, IBM T.J. Watson Thursday, March 25, 2010 4:30pm Small Auditorium, CS105 Emerging trends in computation such as cloud computing, virtualization, and trusted computing require that computation be carried out in remote and hostile environments, where attackers have unprecedented access to the devices, the data and the programs. This poses new problems and challenges for cryptography. In this talk, I will present two such challenges, and my recent work towards solving them. 1. Protecting against Side-channel Attacks: Computing devices leak information to the outside world not just through input-output interaction, but through physical characteristics of computation such as power consumption, timing, and electro-magnetic radiation. Such information leakage betrays information about the secrets stored within the devices, and has been successfully utilized to break many cryptographic algorithms in common use. These attacks are commonly called side-channel attacks. Side-channel attacks are particularly easy to carry out when the device is in the physical proximity of an attacker, as is often the case for modern devices such as smart-cards, TPM chips, mobile phones and laptops. In the first part of the talk, I will describe my recent work that lays the foundation of leakage-resilient cryptography - the design of cryptographic schemes that protect against large classes of side-channel attacks. 2. Computing on Encrypted Data: Security in the setting of cloud computing involves a delicate balance of privacy and functionality: while the client must encrypt its data to keep it private from the server, it should also allow for the server to compute on the encrypted data. Can we simultaneously achieve these opposing goals? In the second part of the talk, I will describe an elementary construction of a cryptographic mechanism that allows computation on encrypted data (also called a fully homomorphic encryption scheme). Both these works leverage new mathematical techniques based on geometric objects called lattices. Vinod Vaikuntanathan is a postdoctoral fellow in the cryptography group at IBM T.J. Watson. He received a Ph.D. from MIT in 2009 under the guidance of Shafi Goldwasser. He is a recipient of the MIT Akamai Graduate Fellowship, the IBM Josef Raviv Postdoctoral Fellowship, and more recently, the MIT George M. Sprowls award for the best Ph.D. thesis in Computer Science. The focus of his research is involves the dual goals of devising new mathematical tools for cryptography, as well as applying theoretical cryptography to counter practical attacks.

1 1

Colloquium Speaker, Wed March 24- David Levin
by Nicole E. Wagenblast 22 Mar '10

22 Mar '10

Systems without Cooperation David Levin, University of Maryland Wednesday, March 24 -4:30pm Small Auditorium, CS105 Successful networked systems must account for potentially competing interests: the Internet is no longer the cooperative, technological playground it once was. The protocols are the rules relegating the venue for this competition, but those rules often lack enforcement. I will present the application of economics and trusted hardware to keep participants in a networked system from deviating from the letter and spirit of a protocol. First, I will apply economic mechanism design to keep selfish users from gaining at the expense of others. I will show that the popular BitTorrent system uses, not tit-for-tat as widely believed, but an auction to decide which peers to serve. This model captures known, performance-improving strategies, and shapes or thinking toward new, effective incentive mechanisms. Second, I will apply trusted hardware to keep both selfish and malicious users from "equivocating," or sending semantically conflicting messages. I will present TrInc (Trusted Incrementer), a small piece of trusted hardware intended for use in large-scale distributed systems. With case studies and an implementation, I will demonstrate that TrInc is a practical primitive for protecting a wide range of systems. These two examples together demonstrate the importance of aligning the assumptions of economics and large-scale systems. Doing so allows us to develop new mechanisms that foster cooperation among the otherwise self-interested.

1 1

A Golovinskiy FPO
by Melissa Lawson 15 Mar '10

15 Mar '10

Alex Golovinskiy will present his FPO on Thursday March 18 at 9:30AM in Room 402. The members of his committee are: Tom Funkhouser, advisor; Szymon Rusinkiewicz and Adam Finkelstein, readers; David Dobkin and Rob Schapire, nonreaders. A copy of his thesis, "Analysis of Global Properties of Shapes", is on display in Room 310. Everyone is invited to attend his talk.

1 0

Distinguished Colloquium Speaker Wed. March 10- Barbara Liskov
by Nicole E. Wagenblast 08 Mar '10

08 Mar '10

The Power of Abstraction Barbara Liskov Wednesday, March 10- 4:30pm Friend Convocation Room Abstraction is at the center of much work in Computer Science. It encompasses finding the right interface for a system as well as finding an effective design for a system implementation. Furthermore, abstraction is the basis for program construction, allowing programs to be built in a modular fashion. This talk will discuss how the abstraction mechanisms we use today came to be, how they are supported in programming languages, and some possible areas for future research.

1 1

Colloquium Speaker, Mon. March 8, 2010- Ranjit Jhala
by Nicole E. Wagenblast 04 Mar '10

04 Mar '10

Algorithmic Software Verification Ranjit Jhala, University of California, San Diego Monday, March 8, 2010- 4:30pm Small Auditorium, CS 105 Software is everywhere, and yet, extremely unreliable. One key to reliability is the design of cost-effective techniques whereby developers can formally specify essential properties of their code, and, machines can rigorously verify that the properties hold, or demonstrate corner cases where they fail. In this talk, we describe two such techniques. First, a scheme that uses logical predicates and theorem provers to automatically compute safety proofs. Second, an approach that uses random walks to find liveness bugs in distributed systems. We describe the key algorithms underlying these approaches and the resulting tools that have helped verify and find subtle defects in a variety of complex software.

1 1

Today's Colloquium moved to CS104
by Nicole E. Wagenblast 03 Mar '10

03 Mar '10

Today's Colloquium moved to the Large Auditorium CS104 Behind the Scenes at Pixar Rob Cook, Pixar Today- 4:30pm LARGE Auditorium CS 104 This talk takes you behind the scenes at Pixar Animation Studios for a look at how its 3D computer graphics films are made. The process starts with the development of the story and continues with modeling the geometry, animating the characters, simulating things like water and cloth and hair, defining the look of the surfaces, putting lights in the scene, and rendering the images. Making a computer animated film requires a close collaboration between artists and technical experts in many areas of expertise and is a great example of the value of bringing different disciplines together. Rob was the co-architect and primary author of RenderMan, software that creates photo-realistic computer images. Every film nominated for a Visual Effects Oscar in the last 15 years has used RenderMan. In 2001, along with two colleagues, Rob received an Oscar for his contributions, the first ever given for software. He has also received the Steven A. Coons Award for lifetime achievement in computer graphics and was inducted into the National Academy of Engineering. Rob has a Bachelor's degree in Physics from Duke University and a Master's in Computer Graphics from Cornell University. At Cornell, he worked on simulating realistic surfaces, taking computer-generated images beyond the distinctive plastic look they had at the time. In 1981, he joined Lucasfilm / Pixar, where he developed the first programmable shader, which is now an essential part of GPUs and game engines. He was the first to use Monte Carlo techniques in computer graphics, which was essential for the simulation of complex, realistic lights and camera effects. The latter proved particularly important in the special effects industry, because it allowed computer-generated imagery to match the motion blur and depth of field of the live-action footage with which it was combined. He is currently the Vice President of Advanced Technology at Pixar.

1 0

Distinguished Colloquium Speaker Wed. March 3- Rob Cook
by Nicole E. Wagenblast 01 Mar '10

01 Mar '10

Behind the Scenes at Pixar Rob Cook, Pixar 4:30pm Small Auditorium CS 105 This talk takes you behind the scenes at Pixar Animation Studios for a look at how its 3D computer graphics films are made. The process starts with the development of the story and continues with modeling the geometry, animating the characters, simulating things like water and cloth and hair, defining the look of the surfaces, putting lights in the scene, and rendering the images. Making a computer animated film requires a close collaboration between artists and technical experts in many areas of expertise and is a great example of the value of bringing different disciplines together. Rob was the co-architect and primary author of RenderMan, software that creates photo-realistic computer images. Every film nominated for a Visual Effects Oscar in the last 15 years has used RenderMan. In 2001, along with two colleagues, Rob received an Oscar for his contributions, the first ever given for software. He has also received the Steven A. Coons Award for lifetime achievement in computer graphics and was inducted into the National Academy of Engineering. Rob has a Bachelor's degree in Physics from Duke University and a Master's in Computer Graphics from Cornell University. At Cornell, he worked on simulating realistic surfaces, taking computer-generated images beyond the distinctive plastic look they had at the time. In 1981, he joined Lucasfilm / Pixar, where he developed the first programmable shader, which is now an essential part of GPUs and game engines. He was the first to use Monte Carlo techniques in computer graphics, which was essential for the simulation of complex, realistic lights and camera effects. The latter proved particularly important in the special effects industry, because it allowed computer-generated imagery to match the motion blur and depth of field of the live-action footage with which it was combined. He is currently the Vice President of Advanced Technology at Pixar.

1 1