Tech that analyzes videos wins top prize at Univ. of Washington computer science showcase
From left: Madrona Prize runner-up winner Jared Nakahara, Madrona partner Chris Picardo, and Madrona Prize winners Dong He, Maureen Daum, and Enhao Zhang. (Madrona Photo / Erika Shaffer)
A student team working on tech to extract information from videos took home the 15th annual Madrona Prize Tuesday at the 2022 Annual Research Showcase at the Paul G. Allen School of Computer Science and Engineering.
The day of presentations, talks and posters culminates in the presentation of the prize to the team that best combines excellent research with commercial potential, awarded by Seattle’s Madrona Venture Group.
The winning team worked on two projects that build on a system called VOCAL that enables users to extract text-based data from large quantities of unlabeled information in videos without pre-trained models. VOCAL (video organization and interactive compositional analytics) enables users to organize their data, build models for specific purposes, and search for events. The two projects were Video Organization and Exploration and Interactive Video Analytics for Compositional Queries, involving search technology.
The projects were led by Ph.D. students Maureen Daum, Enhao Zhang, and Dong He, who were co-advised by Allen School professor and director Magdalena Balazinksa and professor Ranjay Krishna. The team also collaborated with Brandon Hayes, an Allen School graduate now at Microsoft.
Since its founding more than two decades ago, Madrona has funded 21 companies spun out of the UW, including publicly traded Impinj and Turi, which was acquired by Apple. More recently the venture firm backed A-Alpha Bio, Modulus Therapeutics, and OctoML, which won a Madrona Prize five years ago and has raised $132 million in venture capital.
“The University of Washington and the Allen School specifically has grown to be not just a source of incredible innovation for our region but one of the top universities in the world,” said Madrona Managing Director Tim Porter in a statement announcing the award.
The showcase also featured three runner-ups for the Madrona award and two people’s choice winners, noted below.
Madrona Award runner-ups
ClearBuds: Wireless Binaural Earbuds for Machine Learning Based Speech Enhancement. The ClearBuds neural network enables real-time speech enhancement on an iPhone, and suppresses unwanted noises.
Team members: Ishan Chatterjee, Maruchi Kim and Vivek Jayaram, advised by Shyam Gollakota, Ira Kemelmacher-Shlizerman, Shwetak Patel and Steven M. Seitz.
Data Analysis Systems for Statistical Non-Experts. This team is enabling non-experts to generate statistical models through an interactive, no-code process.
Team members: Eunice Jun, advised by Jeffrey Heer and Rene Just.
Levity: Contactless Robotics and Automation for Synthetic Biology. This team is developing tech to suspend and manipulate objects at the millimeter-scale using ultrasound levitation.
Team members: Jared Nakahara, advised by Joshua R. Smith.
People’s choice award winners
Beyond WEIRDNESS of NLP. Team members: Katharina Reinecke, Yulia Tsvetkov, Maarten Sap, Andrew Meltzoff, Ronan Le Bras, Jenny Liang and Rodolfo Barragan.
Distributing Trust and Establishing Transparency in Societal Scale Computing and Network Infrastructure. Team members: Sudheesh Singanamalla, Matthew Johnson, Esther Han Beol Jang, Nick Durand, Abhishek Shah, Spencer Sevilla, Richard Anderson and Kurtis Heimerl.