Sign In

Communications of the ACM

ACM TechNews

Machine-Learning Maestro Michael Jordan on the Delusions of Big Data and Other Huge Engineering Efforts


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
University of California, Berkeley professor Michael I. Jordan.

University of California, Berkeley professor Michael I. Jordan says the rhetoric surrounding machine learning and other major fields of computer science often goes too far.

Credit: Randi Klett

In an interview, University of California, Berkeley professor Michael I. Jordan discusses the ways in which he sees the rhetoric surrounding machine learning and other major computer science fields going too far and making promises it cannot keep.

In particular, Jordan sees "neural realism," the move to metaphorically link many machine-learning efforts to the human brain, as inappropriate and inaccurate, because the current understanding of how the brain works is still too rudimentary to truly be used as a basis for computer systems.

Big data is another field in which Jordan sees extravagance and carelessness. He says big data is being pushed aggressively without any thought for the ways it can go wrong, and he worries about a future swamped by false positives from big data systems. Jordan says the pools of data being used to feed analytic systems are growing so vast that false positives are inevitable, and this inevitability needs to be accounted for with the incorporation of error bars and other statistical methods.

He also believes advances in computer vision are being overhyped; Jordan says the technology is still in its infancy and has a long way to go before it begins to approach the capabilities of human vision.

From IEEE Spectrum
View Full Article

 

Abstracts Copyright © 2014 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account