Undergrads Innovate with Watson Supercomputer for Chance at $100,000
December 28, 2014
L-R: Vincent Tian, Jessie Salas,Vi Tran, Prof. Solomon Darwin, Andrew Koth, David Park, and David Fang.
Team Patent Fox is heading to IBM Watson’s new headquarters in New York City’s Silicon Alley January 9 to test the mettle of a new patent application it developed in the Open Innovation, Leveraging IBM Watson course. The team will vie against nine other teams in the national competition for $100,000.
The winning team includes three business and three engineering majors: Vincent Tian, BS 16, Jessie Salas, BS 16, Vi Tran, BS 15, Andrew Koth, BS 15, David Park, BS 15, and David Fang, BS 16. Patent Fox beat three rival Berkeley-Haas teams Dec. 12 to get the chance to move forward in the IBM competition.
Watson is probably best known for beating the popular Jeopardy quiz show champs. By incorporating three key components—natural language processing, hypothesis generation/evaluation, and dynamic learning—Watson processes information in a way that is more like a human than a computer. It can process over 200 million records per second and learns over time as more information flows into it.
Open Innovation, Leveraging IBM Watson is taught by Solomon Darwin, executive director of the Garwood Center for Corporate Innovation. Berkeley-Haas is one of 10 universities and colleges across North America chosen to offer the course.
Patent Fox designed its app for businesses and law firms. The app relies on Watson’s natural language processing abilities and contextual analysis to help organizations search for patent overlaps more quickly and thoroughly.
The team says the app will reduce the excessive cost and time typically associated with filing patent applications and ultimately help companies protect their patents. The average cost of filing a patent ranges between $1,200 to $6,000. Organizations spend $1.2 billion per year in the U.S. on so-called prior art searches required to prove a new patent’s originality.
Patent Fox developed its plan with the help of a patent attorney and a patent examiner who were already working at a UC Berkeley patent startup, Park says. “We were connected to (the startup) and had a similar idea,” he says. “They were using programmatic computing to solve a problem and we thought if we could use Watson the results would be even better. The way you train Watson is to train it to think and Watson becomes more human in that sense.”
Y. Subramanyam, CEO of Apollo Hospitals, Asia's largest healthcare group, called the team’s idea “just brilliant,” considering the exponential rise in patent filings in Asia.
The runner-up Berkeley-Haas team, Health Note, developed an app that Apollo Hospitals could use to conduct better followup with patients after they leave the hospital. (Health Note will not move on to compete in New York).
The winning team January 9 will receive $50,000 from the IBM Watson Ecosystem group and $50,000 from The Entrepreneur's Fund, a technology venture firm. Students also receive continued access to IBM’s Watson Developer Cloud and become part of the Watson Ecosystem partner program.