Stefan Savage and Pavel Pevzner named 2010 ACM Fellows

Congrats to Stefan Savage and Pavel Pevzner, both named 2010 ACM Fellows.

Pavel Pevzner

Stefan Savage

Stefan Savage is being recognized by the ACM “for contributions to large scale systems and network security.”

Pavel Pevzner is being recognized by the ACM “for contributions to algorithms for genome rearrangements, DNA sequencing, and proteomics.”

Both are professors in the Department of Computer Science and Engineering at UC San Diego.

UC San Diego computer science professor Stefan Savage’s work has focused on the security threats enabled by broad Internet connectivity; including worms, viruses, denial-of-service, botnets and spam. He is known for advancing a quantitative approach towards computer security, including the development of empirical techniques to measure and analyze global-scale attacks and efforts to identify the economics driving modern attackers.

Savage is part of the Systems & Networking and Security research groups in the Department of Computer Science and Engineering at UC San Diego. Savage is also the Acting Director of the UC San Diego Center for Networked Systems. His interests range from the economics of e-crime, to characterizing availability, to automotive systems to routing protocols and data center virtualization.

UC San Diego computer science professor Pavel Pevzner’s work has focused broadly on algorithmic aspects of bioinformatics. Bioinformatics has become a part of modern biology and often dictates new fashions, enables new approaches, and drives further biological developments. Pevzner’s work focuses of DNA sequencing, proteomics, and genome rearrangements, the key areas of algorithmic biology. He authored influential graduate and undegraduate textbooks on bioinformatics algorithms and is now interested in new approaches to interdisciplinary computer science education and introducing computer science to education of biologists. 

Pavel Pevzner is the Ronald R. Taylor Professor of Computer Science at UC San Diego; Director of the NIH Center for Computational Mass Spectrometry; and a Howard Hughes Medical Institute (HHMI) Professor.

What is the state of certain kinds of security problems in the real web?

Another tidbit related to the history sniffing story from computer scientists at UC San Diego.

“The tool we ended up with is useful for answering questions that no one knew how to answer at all before: ‘what is the state of certain kinds of security problems in the real web?’” said computer science professor Hovav Shacham from the Department of Computer Science and Engineering (CSE) at the UC San Diego Jacobs School of Engineering. 

The UC San Diego computer scientists did this while working on the cutting edge of how to efficiently do information tracking in a JavaScript setting.

History Sniffing and Browser Tracking / What about Mobile Browsers?

While the original history sniffing research presented at CCS10  did not look at mobile browsers, the UC San Diego computer scientists say that is something they are looking into now. 

“In general, the mobile browsers and mobile sites seem to be way behind in security,” said computer science professor Hovav Shacham from the Department of Computer Science and Engineering (CSE) at the UC San Diego Jacobs School of Engineering. 

The UC San Diego Jacobs School of Engineering history sniffing press release is here. 

A few more details on the History Sniffing / Browser History / JavaScript story from UC San Diego

Another tidbit on the history sniffing story from the computer scientists here at UC San Diego:

The UC San Diego computer scientists added their new JavaScript information flow engine to the Chrome browser, but the approach could be implemented in other browsers. The history-sniffing detection tool currently adds an additional 60 to 70 percent to total page loading time over a fast network – fast enough to be a useful research tool but too slow for consumer use.

Read the UC San Diego history sniffing press release.

The Forbes story is below: 

History Sniffing: How YouPorn Checks What Other Porn Sites You’ve Visited and Ad Networks Test The Quality of Their Data

FICO, UC San Diego Announce Winners of International Predictive Analytics Competition

SAN DIEGO, Nov 24, 2010 (BUSINESS WIRE) -- FICO , the leading provider of analytics and decision management technology, and the University of California, San Diego (UCSD) today announced the winners of the seventh annual UCSD-FICO Data Mining Contest. Participants from six countries on four continents were among the winners who developed predictive analytics to determine which consumers were most likely to shop online.

This year, contestants were given anonymous data for more than 130,000 consumers. The data included no personally identifiable information. Based on that data, competitors built models to predict which consumers were most likely to buy products online. Participants were judged on how accurately they were able to predict future purchases.

The competition was divided into two categories -- one category utilized raw data, and one category utilized transformed data -- and each category had a Graduate and Undergraduate division. The top three finishers in each category and each division shared $10,000 in cash prizes.

"Students around the world look forward to the UCSD-FICO Data Mining Contest each year as an opportunity to put the skills they've learned in the classroom to use in a rigorous real-world competition," said professor Charles Elkan of UC San Diego Department of Computer Science and Engineering (CSE) within the Jacobs School of Engineering.

"It's been exciting to see the growth of this competition since its inception, with over 140 participants from six continents competing this year. We are proud of all the students who participated, and we appreciate FICO's ongoing commitment to developing and encouraging the next generation of innovators in predictive analytics," said Elkan.   (Read the full press release here.)

2010 Programming Contest

On October 16, 2010, UCSD students participatedin the Fall 2010 UCSD Programming Contest over the course of five hours. 36 students participated and 20 pizzas were consumed. Below is a summary of thecontest, more details (photos and detailed results) are available at: http://cseweb.ucsd.edu/groups/UCSDProgramContest/fa10/results/

The contest consisted of 8 problems, and the top honors go to Elliott Slaughter for solving 5 problems with a combined time of 420 minutes. Junior David Michon placed second, and Haoxi Fang,Van Lam, Andrew Chen, and Matt Der rounded out the top six, each with 4 problems solved. Below is a list of the top fifteen placers in the contest:
1 Elliott Slaughter
2 David Michon
3 Haoxi Fang
4 Van Lam
5 Andrew Chen
6 Matt Der
7 Ellison Bayan
8 Justin Yuhsong Huang
9 Yimai Fang *
10 Rohan Anil
11 Timothy Martin
12 Daniel Price
13 Eric Dorman
14 Yixin Zhu
15 Tim Wheeler
16 Adrian Teng-Amnuay

* unofficial contestant

Thanks go to Mike Dini and The Dini Group for providingthe prizes, food and drink, and support for the contest. Michael Vrable, Mohammad Moghimi, Do-Kyum Kim,Keliang Zhao and Faculty Coach Michael Taylor ran the contest, with help from UCSD Programming Team Co-Captain Justin Huang, CSES President Chris Moghbel, and other folks at CSES. Additional results (and photos) can be viewed here: http://cseweb.ucsd.edu/groups/UCSDProgramContest/fa10/results/

CAPTCHA Project on NPR Weekend Edition

The August 2010 CAPTCHA paper from computer scientists at UC San Diego made it's way to NPR's Weekend Edition on Sunday. Listen to the story, "Spammers Use the Human Touch To Avoic CAPTCHA"

The paper referenced in the NPR story is below:
Re: CAPTCHAs -- Understanding CAPTCHA-Solving from an Economic Context, by Marti Motoyama, Kirill Levchenko, Chris Kanich, Damon McCoy, Geoffrey M. Voelker, and Stefan Savage from the Department of Computer Science and Engineering (CSE) at UC San Diego, published in Proceedings of the USENIX Security Symposium, Washington, D.C., August 2010.


Savage talks big picture about CAPTCHAs and hits one of the findings from the paper:
 

"On the one hand, CAPTCHAs do not keep the bad guys out, but at the same time, they actually are effective at keeping the problem in control," Savage says.


Below is the abstract to the paper Re: CAPTCHAs -- Understanding CAPTCHA-Solving from an Economic Context:
Reverse Turing tests, or CAPTCHAs, have become an ubiquitous defense used to protect open Web resources from being exploited at scale. An effective CAPTCHA resists existing mechanistic software solving, yet can be solved with high probability by a human being. In response, a robust solving ecosystem has emerged, reselling both automated solving technology and realtime human labor to bypass these protections. Thus, CAPTCHAs can increasingly be understood and evaluated in purely economic terms; the market price of a solution vs the monetizable value of the asset being protected.We examine the market-side of this question in depth, analyzing the behavior and dynamics of CAPTCHA-solving service providers, their price performance, and the underlying labor markets driving this economy.

YY Zhou / More Reliable Computing

Calit2's Tiffany Fox wrote a nice profile on Yuanyuan (YY) Zhou, a computer science professor here at the Jacobs School of Engineering.

“People always ask why their systems fail, and the computer industry is starting to pay attention to this reliability issue,” explains Zhou, who joined the UCSD Department of Computer Science and Engineering (CSE) in the Jacobs School of Engineering in the summer of 2009. “Fundamentally, my research is about making computer systems less vulnerable to attacks so they crash less. When Windows crashes, it’s more of an inconvenience, but when an e-commerce site crashes it can really be a problem.”


Zhou is the first holder of the Qualcomm Endowed Chair in Mobile Computing. The chair is one of four established in the Jacobs School of Engineering through Qualcomm’s original $15 million commitment to the California Institute for Telecommunications and Information Technology (Calit2). Read the story on the Calit2 news site.

Check out a few more great photos of computer sicnece professor YY Zhou taken by Erik Jepsen on the Calit2 Flickr site.

Pedestrian Remover in ScienceNews

Last Friday, ScienceNews covered the Pedestrian Remover story.

I liked the final quote from computer science professor Serge Belongie:

“The project is a nice combination of computer science and social awareness,” he says. “All too often, computer scientists implement something just because it is cool rather than there being a need.”

Read the full story by Rachel Ehrenberg on the ScienceNews site: "Protecting innocent — and not so innocent — bystanders "

Gizmag Glances at SleepServer

Gizmag has a SleepServer story up on their super-high-traffic tech blog today.

Check it out...and the easy-to-read summary from the Jacobs School of Engineering is here, and a longer, more technical summary is here.

Pictures!

Stout: An Adaptive Interface to Scalable Cloud Storage

I just noticed that SleepServer is not the only UC San Diego computer science paper being presented this week at the 2010 USENIX Annual Technical Conference in Boston.

"Stout: An Adaptive Interface to Scalable Cloud Storage," John C. McCullough, University of California, San Diego; John Dunagan and Alec Wolman, Microsoft Research, Redmond; Alex C. Snoeren, University of California, San Diego.

The abstract is below:

Many of today's applications are delivered as scalable, multi-tier services deployed in large data centers. These services frequently leverage shared, scale-out, key-value storage layers that can deliver low latency under light workloads, but may exhibit significant queuing delay and even dropped requests under high load. Stout is a system that helps these applications adapt to variation in storage-layer performance by treating scalable key-value storage as a shared resource requiring congestion control. Under light workloads, applications using Stout send requests to the store immediately, minimizing delay. Under heavy workloads, Stout automatically batches the application's requests together before sending them to the store, resulting in higher throughput and preventing queuing delay. We show experimentally that Stout's adaptation algorithm converges to an appropriate batch size for workloads that require the batch size to vary by over two orders of magnitude. Compared to a non-adaptive strategy optimized for throughput, Stout delivers over 34x lower latency under light workloads; compared to a non-adaptive strategy optimized for latency, Stout can scale to over 3x as many requests.

PCs Sleep (and Work) Through the Night with SleepServer

Leaving your computer on all night/weekend/break just in case you need to connect remotely…that’s SO last decade. UC San Diego computer scientists created software that lets you put your computer to sleep without worrying that you’ll have to schlep your way to campus on Saturday night just because you suddenly need to access your computer connected to an enterprise network. SleepServer wakes up your sleeping PC remotely when you need it…but it won’t wake up the PC if the PC’s alter ego – a lightweight virtual image – can handle the job.

That’s the simplest example of how SleepServer helps UC San Diego cut its carbon footprint. But SleepServer does much more. That sleeping computer of yours also remains active on voice over IP/IM/peer-to-peer networks as well thanks to SleepServer.

How is this possible? When you put your computer to sleep, the software activates a lightweight virtual image of your PC which runs on a commodity server, along with hundreds of images of other PCs on your network. That virtual image can do a number of basic things that would otherwise require the actual PC to be awake.

SleepServer reduces energy consumption on enterprise PCs previously running 24/7 by an average of 60 percent, according to a new study presented today at the 2010 USENIX Annual Technical Conference.

Yuvraj Agarwal from UC San Diego is presenting this work TOMORROW June 25 in Boston at the 2010 USENIX Annual Technical Conference.

Read the story here.

More photos at this Jacobs School blog post.

Wake Up! SleepServer is Here

UC San Diego computer scientists are presenting a new “green computing” paper on Thursday June 24 in Boston at the 2010 USENIX Annual Technical Conference. (read the full story here and check out photos here)

The paper presents the latest on a UC San Diego technology that enables people who usually keep their work computers on all night and over the weekend to put them in low-power sleep mode before they go home. Also, SleepServer allows people to maintain an active presence on voice over IP, IM and peer to peer networks.

Over a two week period last September, 30 people in the computer science building at UC San Diego saved 60 percent energy, on average, compared to leaving the computer on 24/7 (which is what many of the people were doing).

UC San Diego Research Scientist Yuvraj Agarwal from the Department of Computer Science and Engineering developed SleepServer, and its technological predecessor Somniloquy. (Agarwal earned his PhD in computer science from UC San Diego in 2009).

There are related solutions out there, but according to Agarwal, none of the other solutions are able to run pared down versions of applications such as Skype in order to maintain network presence even when the computer is in sleep mode.

“SleepServer: A Software-Only Approach for Reducing the Energy Consumption of PCs within Enterprise Environments” by Yuvraj Agarwal, Stefan Savage and Rajesh Gupta from the Department of Computer Science and Engineering (CSE) at the UC San Diego Jacobs School of Engineering.

In Proceedings of the USENIX Annual Technical Conference (USENIX ATC '10), June 2010.

Yuvraj Agarwal Photos

Photos below for the new story:
Enterprise PCs Work While They Sleep, Saving Energy and Money

Caption info for the graphs is available in the PDF of the SleepServer paper. The photo credit for any of the portraits of Yuvraj Agarwal is: UC San Diego / Erik Jepsen

CSE168: Rending Algorithms First, Second and Honorable Menions...Grand Prize Winner to Come

Update July 2: here is the grand prize image for CSE168. The 2010 CSE 168 Rendering Algorithms press release is here. The first, second and some of the honorable mention images are below. These are all images that computer science students created using rendering algorithms and C++ (not photoshop/illustrator/etc)

(Photos from the competiton are here.)

Update Friday June 11: last two images are new.

First Place: William George and Robert Gross




Second Place: Jason Obenberger




Honorable Mention: Jennifer Chandler



Honorable Mention: Steven Upham



Honorable Mention: Tim Etler



Honorable Mention: Oleg Bisker




Update Friday June 11: Two new honorable mention images below.

Honorable Mention: Joel Chelliah





Honorable Mention: Sid Vijay

Making Graphics with Rendering Algorithms: Class and Contest

I popped into the final presentations for CSE 168 (Rendering Algorithms) taught by computer science professor Henrik Wann Jensen. I'm waiting for the final results, but in the meantime, here are some of the photos from the class. Above: UC San Diego senior Nick Echols shows off a "behind the scenes" graphic tied to his CSE 168 project.

The C++ computer code in the image (below) highlights the fact that these images are created by computer science students who write and manipulate lines of C++ code. This is not about using software that has already been created. It's much more involved than learning Photoshop, Illustrator and 3D graphics programs.