Teaching Machines to Learn by Themselves

Tuesday, Feb 20, 2018

Picture yourself trying to build a machine to detect email spam. You might start with simple rules that identify key words such as “drugs,” for instance. Of course, some legitimate emails could contain those key words, so you add rules to take into account these cases. In an adaptive world, advertisers may quickly learn the rules of your system and develop ways to fool your spam detector. 

Painstakingly identifying patterns and continuously updating your system by hand is a daunting challenge — even more so for tasks such as identifying one human face out of more than 7 billion, or driving a car through chaotic Manhattan traffic. 

But what if you could set an environment and ground rules for machines to teach themselves and constantly improve their performance as more data rolls in?  Now you’ve entered the world of Elad Hazan.

Read More on the Princeton Computer Science website.