Hidden Gems Are Inside UW Computer Science & Engineering. Can They Be Mined?


Although an attentive observer can see a steady stream of news articles from UW Computer Science & Engineering, with more than 40 faculty members and 750 students, there is more going on than is reported. When I was given an opportunity to occupy an office inside the Paul G. Allen Center, I was excited to take a closer look.

One of the most visible gems in CSE is Oren Etzioni. Etzioni has been the technical wizard behind at least four commercial products or services. Most famously, he was involved with Farecast, a web service to help people to decide when to buy airplane tickets at the best price. Farecast was sold to Microsoft and now is a key part of Bing Travel. Etzioni was the technologist behind Go2Net, Netbot, and MetaCrawler, all cutting-edge web services that became commercial products. Now his latest project, Decide.com, just went live, a service for deciding the best time to purchase consumer electronics equipment. The core technology behind Farecast and Decide.com is his approach to indexing and analyzing very large collections of text documents, searching for information to help people make better decisions.

Shwetak Patel is another sparkling gem in CSE. Patel, who specializes in sensor technology, was in the news recently because his company Zensi was sold to Belkin, a maker of computer and consumer electronics products. Through Zensi, Patel commercialized signal processing techniques for measuring electricity, water, and natural gas usage within a house without having to instrument every appliance. One sensor per house is enough to identify what device is using electricity, when it is using it, and how much is being used. The same techniques with different sensors work for water and natural gas. Patel also made a Hollywood appearance by selling a system for detecting the illegal copying of movies by digital cameras in movie theaters, which he sold to the Motion Picture Association of America. Currently he is doing compelling work with networks of low-power sensors that can be easily and cheaply deployed for residential and commercial building monitoring. Patel’s sensors are hidden inside a wall or under a water heater to detect and alert a homeowner, or an insurance company, when a sensor detects water leaks, fire or carbon monoxide. These sensors will be able to save the lives of many homeowners, and save insurers hundreds of millions of dollars; water damage claims now surpass fire claims and natural disasters and are very expensive to settle.

The UW CSE computer security team recently won gleaming accolades at the national collegiate cyber defense competition in April, evidence of the world-class expertise in this area at the UW. The challenge for teams from colleges across the company was to take control of and run an IT system while defending it from attack by professional “hackers” over a three day contest, and the UW team won the competition. Yoshi Kohno, a UW gem focused on computer security, has made news recently for exposing security holes in equipment we use every day. Kohno’s work highlights the potential for hackers to gain control of automobile systems, pacemakers, or baby monitors. Kohno not only exposes security holes, but develops tools and techniques for secure computing.

During my short time in CSE, I have discovered additional diverse and interesting research in several areas of computer science with real possibilities for commercialization.

A group of database experts, led by Magda Balazinska, is studying tools and techniques for creating and managing extremely large databases and storing huge streams of data in real time. The size and amount of data that can be collected today can be overwhelming, especially if the data needs to be organized in a way that allows useful questions to be asked about it. This is deep technical work that underpins future database products.

Faculty and students including Yaw Anokwa and Carl Hartung are working on applying information technologies in developing regions of the world where workers are low-skilled, funding is scarce, and the IT environment is harsh. Anokwa and Hartung use smart phones as data collection devices and push data into the cloud for storage, analysis, and reporting. Their use of these technologies allows health workers to collect medical data in the field and send it to urban centers for analysis. The underlying technology they devised for quickly building data collection applications has been widely adopted by others – the open source library Open Data Kit, developed in collaboration with Google Seattle.

One of several people working on machine learning is Pedro Domingos. One of Domingos’ projects is figuring out how valuable a person is to a marketer based on that person’s social network and their perceived expertise on a given product or subject. Knowing who is best-connected and most expert can allow precise targeting of marketing messages which can have a big impact on marketing effectiveness.

I met a researcher named Justin Cappos who is working on an alternative to cloud computing by creating a pool of shared personal computers that can work together to solve problems. He’s creating a network of safe “netbots” available to each participant in the network. Unlike the SETI network that uses volunteers’ computers for the specific purpose of analyzing collected signal data for evidence of extra-terrestrial life, Cappos’ network lets any participant use all the other computers in the network to help solve computational problems.

Barbara Mones, Director of the Animation Research Lab and formerly of DreamWorks and Industrial Light & Magic, is working on tools to support computer animation. Her students produce compelling, technically advanced animated films. Her recent work is on new tools for planning animations using a real-time storyboard application that allows the animators to work out all the elements of a scene quickly before they spend the time required to fully render each scene.

During my four months on campus I’ve covered less than a quarter of the faculty. I know there are many more gems out there, and the entrepreneur inside of me is excited about the potential to mine them and share them with the world.

My expectation was that technology comes out of academia through its faculty members. News articles report professors developing new technologies and forming companies. But not many professors will leave their university to play a significant operating role in a company intended to commercialize their technology. Although Etzioni and Patel have created commercially successful products, they continue as UW professors, not employees of the companies created to bring their research to market. Their strengths and passions, as with almost all university professors, are research and teaching, not building companies.

More likely, students develop an idea, working under the tutelage of a faculty advisor, and leave the university to form a company to commercialize their work. Students are the entrepreneurs, guided by faculty and other mentors. For example, Etzioni’s four Decide.com co-founders were recent UW undergraduates (three from CSE, one from another program); the company’s chief scientist and other two Ph.D.s are CSE alums; the CEO and 80 percent of the employees are UW alums.

That is the way it ought to be. Faculty members educate students. Faculty members are moving on to the next big idea while students are commercializing the last one. The history of successful university spin-outs in computer science is pretty clear. Google, Yahoo!, Netscape, Sun Microsystems, and many others were started by students.

As chair of the Technology Alliance, I care deeply about our universities and their impact on the Washington state economy. The real gems at the university are most effectively transferred outside by a motivated student who takes an idea and runs with it. Entrepreneurs can be helpful to the institutions, faculty, and students to better understand how to effectively commercialize research results, and to help identify results with commercial potential. Together we can create a culture that values and informs the commercialization of ideas, always remembering that faculty members and outsiders are not often the miners of these gems. Students are.

Jeremy Jaech is CEO of SNUPI Technologies, as well as a serial technology entrepreneur who co-founded Visio and Aldus. Follow @

Trending on Xconomy