The expertise and growth of University of Washington’s computer science department was on full display Wednesday evening on campus in Seattle.
The evening showcase event is a science fair on steroids, with UW students showing off their latest research projects to colleagues, friends, and others from the local tech community.
Walking around the Allen School was like taking a step into the future, where robotics, computer vision, machine learning, data science, networking, and other innovative technologies were demoed.
Seattle-based venture capital firm Madrona Venture Group has awarded the Madrona Prize for the past 11 years to the research with the greatest commercial potential. This year’s top prize ($3,000) went to two groups — here are descriptions of each project with the students and professors who developed them:
- Lab in the Wild, a large-scale online experimentation with diverse uncompensated samples (Allen School postdoc Nigini Abilio Oliveira; Allen School Ph.D. student Eunice Jun; and Allen School professor Katharina Reinecke)
- Augury, a platform that predicts visual appeal of website design (visiting scientist Manuel Nordhoff and Allen School professor Katharina Reinecke)
Lab in the Wild is a marketplace of sorts that collects research from participants around the world on human-computer interaction. The platform has racked up more than four million users, and the resulting data is meant to help compare website preferences of people from different countries, for example, or show developers how to tweak user interfaces for certain age groups with different mouse interaction abilities.
Here’s a bit from the project’s about page:
Our previous experiments on Lab in the Wild have seen several thousands of participants, helping us to shed light on the impact of culture on users’ preferences, motor abilities, perception, and the like. But Lab in the Wild doesn’t just help us answer our questions. It also provides participants with personalized feedback, which lets them compare themselves and their performance to people of other countries.
Augury builds off a similar idea and lets developers do massively-scalable A/B testing for website design.
Madrona Venture Group Managing Director Tim Porter noted how both projects were tightly related.
“The problem that Augury and Lab in the Wild are broadly solving around personalization and website optimization for different demographics — and not just the U.S. audience — is super applicable,” he told GeekWire. “The commercial applications of this network of people is also super exciting.”
Porter said the event is one of Madrona’s favorite nights of the year. The firm has invested in 17 startups that have spun out of the UW’s highly-regarded computer science school. Porter noted how important the school is to Seattle’s tech ecosystem and vice versa; many companies like Microsoft, Tableau, Adobe, and Nutanix worked with students on the projects on display Wednesday.
“It’s a total virtuous cycle,” he added. “One reason the ecosystem has grown is because of UW CSE — and as the ecosystem grows, UW CSE benefits. They both drive each other and that’s exciting.”
The showcase event was also a good example of how the department continues to attract not only top students, but professors and researchers. For example, Reinecke, involved with both of the winning projects, arrived on campus two years ago from the University of Michigan. Manuel Nordhoff, who helped build Augury, is a visiting scientist who was most recently at Cologne University of Applied Sciences in Germany.
Earlier in the day, Siddhartha “Sidd” Srinivasa gave a keynote address. The robotics expert joined the UW this fall after 18 years at Carnegie Mellon University — he also brought along his entire team of more than a dozen undergraduate and graduate researchers, becoming the latest renowned computer scientist to join the UW faculty.
Asked why he came to the UW, Srinivasa cited the ability to collaborate at the university not only with experts in computer science and electrical engineering but also with leading researchers from the medical school, biomechanical engineering and other relevant specialities who are based a short distance away — a walk across campus vs. a flight to another city.
In the decade since Microsoft co-founder Paul Allen helped open the current UW CSE building, the demand for computer science graduates has skyrocketed in the Seattle region, thanks to a strong startup ecosystem, Amazon’s rapid growth, and the opening of Seattle-area engineering offices by Google, Facebook, and many other tech companies based outside the region. More than two-thirds of UW CSE graduates remain in-state after completing their degree.
In 2016, CSE became the leading “first-choice” major among confirmed incoming UW freshmen, surpassing the longtime leading preferred major, Business Administration. The department will open a new building in 2019 — funded in part by Microsoft, Amazon, Zillow, and Google — to help meet the growing demand.
Madrona also gave cash prizes to three runner-up projects:
- A Unified Approach to Interpreting Model Predictions (Allen School Ph.D. student Scott Lundberg and Allen School professor Su-In Lee)
- TVM: End-to-End IR Stack for Deep Learning Systems (Allen School Ph.D. students Tianqi Chen, Thierry Moreau, Haichen Shen; Ziheng Jiang; Allen School professors Luis Ceze, Carlos Guestrin, and Arvind Krishnamurthy)
- Real-Time VR Video Processing with the Hardware-Friendly Bilateral (Allen School Ph.D. student Amrita Mazumdar; Allen School post doc Armin Alaghi; and Allen School professors Luis Ceze and Mark Oskin)
There were also three groups awarded with the “People’s Choice” honor based on votes from attendees at the event:
- Winner: Learning Stylized Character Expressions from Humans (Allen School Ph.D. students Deepali Aneja and Bindita Chaudhuri; Alex Colburn; Gary Faigin; and Allen School professors Linda G Shapiro and Barbara Mones)
- Runner up: RoyalFlush: Non-invasive toilet water overflow detector (EE Ph.D. students Farshid Salemi Parizi and Josh Fromm; Allen School professor Shwetak Patel)
- Runner up: IDCam: Precise Item Identification for AR-Enhanced Object Interactions (Allen School Ph.D. students Hanchuan Li, Eric Whitmire, Alex Mariakakis, and Victor Chan; Alanson Sample; and Allen School professor Shwetak Patel)