Software engineer Tracy Chou observes how bias is easier to remove from machine-learning models than from human behavior. But she warns that bias can still creep in through the use of proxy data: for instance, excluding data on race but including closely related information like zip codes. An early employee of several Silicon Valley startups, Chou was a founding advisor to Project Include and continues to push for more diversity in tech.
Related
Luke Sykora,
Stanford University
Starting Up in a Downturn
An Entrepreneurial Thought Leaders mini-guide for aspiring founders facing a troubled economy.
Josh Makower, MD,
Stanford Byers Center for Biodesign
The Biodesign Innovation Process [Entire Talk]
Innovation isn’t random – it’s a process that can be learned, improved, and effectively deployed to solve specific problems.
Josh Makower, MD,
Stanford Byers Center for Biodesign
The Biodesign Innovation Process [Entire Talk]
Innovation isn’t random – it’s a process that can be learned, improved, and effectively deployed to solve specific problems.
John Felts,
Cruz Foam
Engineering Green Materials [Entire Talk]
To scale a technology, engineers need to think beyond the technology itself.
John Felts,
Cruz Foam
Engineering Green Materials [Entire Talk]
To scale a technology, engineers need to think beyond the technology itself.
