Allen School News

Researchers in UW CSE’s Security & Privacy Research Lab turned archaeologists to deliver the first comprehensive study of third-party web tracking based on a new tool, TrackingExcavator, that detects and analyzes third-party tracking behavior. UW CSE Ph.D. student Adam Lerner presented the results of the study, which examines tracking on the most popular online destinations dating back to 1996, at the USENIX Security Conference in Austin, Texas last week.

“Third-party tracking started quite early in the history of the web,” Lerner noted in a UW News release. “People are becoming more concerned about the potential impact of third-party web tracking, but we lacked a comprehensive history of how trackers — and the types of information they collect — have evolved over time.”

The team, which in addition to Lerner includes CSE Ph.D. student Anna Kornfeld Simpson and CSE professors Franzi Roesner and Yoshi Kohno, set out to build that history by reconstructing tracking data for the top 500 websites using web pages archived in the Wayback Machine. The task was made more complicated by the fact that no one anticipated, when putting together those early websites, that we would want or need to trace the evolution of third-party tracking decades later.

“Reconstructing tracking behavior from the Wayback Machine is difficult because it was designed to archive web content, not tracking techniques,” Kornfeld Simpson told UW News. “We had to develop techniques to extract tracking information from the archive. For example, we collected tracking cookies from archived HTTP headers and Javascript and then simulated the browser’s cookie storage behaviors to detect tracking behavior.”

They found that activity on popular websites by third-party trackers—such as advertisers, analytics engines and social media widgets—has increased four-fold over the past two decades. Tracking has also become more complex, evolving from simple cookies and pop-up windows to more sophisticated methods.

According to the news release,

“Today, the average top website has an average of at least four third-party trackers looking at user activity. The team stresses that these numbers are likely underestimates, since not all websites are fully archived.

“They also found that today individual trackers cover a much larger fraction of the web.…These findings are important to understanding the effects of tracking on privacy, since tracking users on more sites allows trackers to develop a more detailed and intimate picture of their behavior.

“This 20-year historical perspective paints a clear picture of how third-party tracking has evolved with the rise and fall of different techniques, advances in technology, and our increasing reliance on the web in our lives. In general, third parties are watching and collecting information. How we may feel about that remains to be seen.”

Read the complete UW News release here, and the research paper here. Learn more and gain access to the team’s data on the TrackingExcavator website here. Read coverage of the study in TechCrunch, USA Today, Fortune and IEEE Spectrum, and watch video from NBC Today and KOMO News.

The project is the latest example of UW CSE’s leadership in web privacy research, including previous work by Roesner, Kohno, and then-CSE professor David Weatherall to analyze and classify web-tracker behavior and to empower users with tools such as ShareMeNot, which was subsequently incorporated into the Electronic Frontier Foundation’s Privacy Badger. Read more →

Sarah Stone

Micaela Parker

The careers of Micaela Parker and Sarah Stone, program managers at the UW eScience Institute, are the subject of an article in the latest issue of Oceanography, the quarterly magazine of The Oceanography Society. In it, Micaela and Sara talk about what inspired them to jointly apply to the program manager position, in which they share responsibility for day-to-day operations of the institute and serve as the primary point of contact for campus and external partners and the public.

“The eScience Institute’s mission is to engage researchers across disciplines in developing and applying advanced computational methods and tools to real world problems in data-driven science and research,” they explain. “This interdisciplinary mission attracted us to the program manager position.”

Micaela and Sarah proposed a job-sharing arrangement—backed up by data, of course!—that would allow them to balance a career in academia with their roles as caregivers. The arrangement has worked out well for all parties, as it has enabled the faculty, staff and students that work with the institute to benefit from both women’s skills and experience.

“We enjoy helping scientists create and utilize tools that enable novel research questions and empower others,” Micaela and Sarah tell Oceanography. “The eScience Institute is an exciting, cross-disciplinary environment that allows us to work with scientists doing cutting-edge research.”

They credit the institute’s leadership, including UW CSE professor Ed Lazowska and iSchool professor and CSE adjunct professor Bill Howe, plus Chris Cunnington and the rest of the UW CSE staff, for supporting their job-share. They offer advice to other women in academia similarly striving for work-life balance, urging them to not be afraid seek flexibility but also to embrace “opportunities to explore outside of your comfort zone.”

Read a PDF of the article here. Read more →

Inspired by the global nature of the Olympic games, the UW CSE Systems Group’s Icelandic Ph.D. students treated us to a lunch of treats from their homeland: fermented shark, smoked and salted foal meat, sheep’s head cheese, liver sausage, blood pudding, flat bread with smoked lamb and butter, sheep pate on rye bread, …

Michael Phelps may have won twenty two gold medals, but we bet he wouldn’t have been able to work his way through this stuff … Read more →

Researchers in the UW Interactive Data Lab led by UW CSE professor Jeffrey Heer have captured the Best Paper Award at InfoVis 2016 for their paper presenting Vega-Lite, a new high-level grammar for rapid and concise specification of interactive data visualizations.

Although interaction is at the heart of effective data visualization, existing systems for designing interactive visualizations are either complex or overly limiting—particularly when it comes to customization. Enter Vega-Lite, which combines a traditional grammar of graphics—including visual encoding rules and a composition algebra for layered and multi-view displays—with a novel interaction grammar to bring the advantages of high-level specification to interactive visualization.

The research team, which includes Stanford Ph.D. student Arvind Satyanarayan and UW CSE Ph.D. students Dominik Moritz and Kanit Wongsuphasawat, demonstrated the software’s effectiveness in conjunction with common visualization techniques, such as panning, zooming and linked selection, as well as customized interaction methods. Using Vega-Lite, analysts will be able to produce and modify interactive graphics with the same ease as they now construct static plots.

From the award citation:

“Vega-Lite is a high-level visualization grammar that integrates an algebra for interaction techniques with operations on views. An important goal of this work is to define a high-level language in which to specify sophisticated interactive visualizations that can be generated automatically. High-level languages like these may help non-programmers to create interactive visualizations and ease them into specifying visualizations in a programmatic way.”

The team will present Vega-Lite at InfoVis, which is part of the big IEEE VIS 2016 conference, in Baltimore, Maryland in October.

Learn more about Vega-Lite here, and read the research paper here.

Way to go, team! Read more →

UW CSE professor Maya Cakmak and her efforts to train robots to perform everyday tasks are the subject of a recent MIT Technology Review article that posed the question, “What will it take to get a robot to clean your home so you don’t have to do it?”

As it turns out, we have a ways to go before you will be able to trade your Roomba for a Rosie.

Cakmak specializes in programming by demonstration, in which a robot is trained to perform a task by watching and imitating a human’s performance of the same task. The goal is to enable the robot to generalize what it learns through the demonstration—such as a particular cleaning technique—so that it can apply that technique using different tools for different situations. But as Cakmak notes, “Cleaning is different from other tasks we’ve thought about in robotics, which [have] typically involved manipulating objects, or moving them place to place.” For example, a robot would have to learn to identify which tool is the right one for a job, and the correct speed and pressure to apply while using it.

These and other challenges Cakmak identified—such as a lack of machine-friendly design when it comes to many current dwellings—will have to be addressed before autonomous household robots move out of the realm of science fiction and into our homes.

Plus, we need robots with waterproof appendages because vacuuming is a pleasure compared to cleaning the toilet.

Check out the full article here. Watch a video demonstration here. Read more →

GeekWire is reporting that Turi, the spin-off founded by Carlos Guestrin, Amazon Professor of Machine Learning in UW CSE, has been acquired by Apple.

According to GeekWire’s breaking news post:

“Machine learning and artificial intelligence startup Turi has been acquired by Apple in a deal characterized as a blockbuster exit for the Seattle-based company, formerly known as Dato and GraphLab, GeekWire has learned.

“The acquisition reflects a larger push by Apple into artificial intelligence and machine learning. It also promises to further increase the Cupertino, Calif.-based company’s presence in the Seattle region, where Apple has been building an engineering outpost for the past two years …

“Sources close to the deal pegged the purchase price at around $200 million …”

On July 21 we reported on the IPO of UW CSE spin-off Impinj, whose current market cap is roughly $350 million. While representatives of Apple and Turi are keeping mum, if we take today’s GeekWire report at face value, this marks two great exits of UW CSE companies – both backed by our friends at Madrona Venture Group – in two weeks!

Read the full article here.

More from GeekWire: “Why Apple bought Turi.” Read more →

Luna Dong

Magda Balazinska

Professor Magda Balazinska  and Ph.D. alum Luna Dong of UW CSE’s Database group will be honored for their research contributions to the field of data management at the Very Large Data Bases (VLDB 2016) conference next month. Balazinska receives the inaugural VLDB Women in Database Research Award for her “inspirational research record on scalable distributed data systems.” Dong, who earned her Ph.D. in 2007 working with former UW CSE professor and Database group founder Alon Halevy, receives the VLDB Early Career Research Contribution Award for “advancing the state of the art of knowledge fusion.”

Balazinska holds the Jean-Loup Baer Career Development Professorship in UW CSE and is a Senior Data Science Fellow at the UW’s eScience Institute. Her research focuses on big data management and scientific data management in cloud computing environments. In addition to co-leading the Database group, she co-founded the UW’s AstroDB group—a collaboration between UW CSE, the eScience Institute, and the Astronomy department—to develop novel tools and techniques for processing, storing and querying the massive volumes of data being generated by telescopes and simulations. Her projects extend beyond astronomy and include, among others, collaborations with neuroscientists and biologists. Balazinska is one of the leaders of the Myria project, which develops and operates a fast, flexible cloud-based service for big data management and analytics. A demonstration version of the service is available through the project website. Prior to her work on Myria, she led the Nuage project, which enabled scientists to store and analyze large volumes of data using the Hadoop system. She also worked on the SciDB multidimensional parallel array engine.

Balazinska has earned numerous honors for her work, including an NSF CAREER Award, a Microsoft Research New Faculty Fellowship, a Google Research Award, an HP Labs Research Innovation Award, and multiple best paper awards. In addition to her research and teaching, Balazinska serves as director of the IGERT Ph.D. program in Big Data and Data Science and co-founded the Northwest Database Society (NWDB), which brings together researchers and practitioners in the database and data management fields from across the Pacific Northwest.

Dong’s research interests span knowledge management, data integration and quality, and applied machine learning. After graduating from UW CSE, she spent five years at AT&T Research, where she helped build the company’s Data Fusion research area and systems for gauging the trustworthiness of data sources and the best sources for data integration.

In 2013, Dong became a senior research scientist at Google, where she initiated the Knowledge Fusion research area and contributed to the company’s Knowledge Vault project, which automatically fuses the knowledge collected from billions of web pages into a probabilistic knowledge base. Dong also invented Knowledge-Based Trust, a system that gauges the trustworthiness of web sources and aids the verification of facts—referred to as the “Google Truth Machine” in a Washington Post article about the project. Recently, Dong joined Amazon as a principal scientist, initiating and leading the effort to build Amazon Product Graph.

VLDB is one of the two top database conferences in the world (the other one being SIGMOD). VLDB 2016 will be held September 5-9 in New Delhi, India.

Congratulations, Magda and Luna! Read more →

UW CSE Ph.D. alumnus Gary Kildall was a pioneer of personal computer software. He wrote programming language tools including assemblers (Intel 4004), interpreters (BASIC), and compilers (PL/M). He created a widely-used disk operating system (CP/M). He and his wife, Dorothy McEwen, started a successful company called Digital Research to develop and market CP/M, which for years was the dominant operating system for personal microcomputers.

Gary died in 1994, at the young age of 52. In 1993, the year before his death, he wrote a draft of a memoir titled “Computer Connections: People, Places, and Events in the Evolution of the Personal Computer Industry.” Gary’s children, Scott and Kristin, have made the first portion of that memoir, along with their introduction to it and previously unpublished family photos, available via the Computer History Museum.

Scott and Kristin write: “Gary viewed computers as learning tools rather than profit engines. His career choices reflect a different definition of success, where innovation means sharing ideas, letting passion drive your work and making source code available for others to build upon. His work ethic during the 1970s resembles that of the open-source community today.”

Read the preamble and the remarkable manuscript here.

UW CSE is proud to count Gary as one of our most distinguished alumni: Ph.D. number 7 from our department, awarded in 1972, advised by Hellmut Golde. Read more →

UW CSE professor and alum Tom Anderson (Ph.D., ’91) and alum Albert Greenberg (Ph.D., ’83) have been elected to the Washington State Academy of Sciences. Anderson and Greenberg, who were elected to the National Academy of Engineering earlier this year, are among two dozen new WSAS members selected based on “their outstanding record of scientific achievement and willingness to work on behalf of the academy in bringing the best available science to bear on issues within the state of Washington.”

The WSAS was created by the Washington State Legislature in 2005 to harness the expertise of leading scientists and engineers to provide authoritative and objective scientific and technical analysis to public policy-makers and to increase the role and visibility of science in the state, modeled after the National Academies’ National Research Council. Anderson and Greenberg join UW CSE professors Ed Lazowska and Hank Levy, adjunct professors David Baker, Tom Daniel and Joseph Felsenstein, and affiliate faculty members Phil Bernstein, Eric Horvitz, and Burton Smith of Microsoft Research, Lee Hood of the Institute for Systems Biology, Radia Perlman of EMC, and Rick Szeliski of Facebook as members. They will be formally inducted at the WSAS annual meeting in September.

Read our previous blog post on Anderson and Greenberg here.

Congratulations, again, to Tom and Albert on their tremendous achievements! Read more →

Impinj, the Seattle-based RFID company founded by UW CSE professor Chris Diorio and his Caltech Ph.D. advisor Carver Mead, made its debut on Wall Street today to great fanfare. Impinj is the first Seattle technology company to go public in 2016.

From the GeekWire article:

The Impinj team rang the closing bell at NASDAQ today to celebrate their IPO

“Wall Street likes what they see in Impinj, a 16-year-old Seattle-based maker of Radio Frequency Identification technology that today went public on Nasdaq at $14 per share. That was the upper end of the range for the company, which makes RFID chips that allow retailers to track inventory or manufacturers to track parts….

“Impinj, which is trading under the ticker PI, is doing well in its debut. The stock shot up more than 20 percent, and it is now trading around $17.17.” (Impinj closed the day at $17.97, up 28% from the $14.00 IPO price.)

Congratulations to Chris, Carver and the entire Impinj team! And also to the investors who have stood by the company – particularly our good friends at Madrona Venture Group, which has backed more than a dozen UW CSE startups.

Read all about it on GeekWire here, TechCrunch here, and Xconomy here. Read more →