What's the difference between artificial intelligence, machine learning, deep learning and neural networks? Don't trust vendors' marketing materials to help you find a workable, accurate definition.
"It's almost criminal, the use of these terms by the vendors today," says Kris Lovejoy, CEO of security firm BluVector. "Pretty much anybody that has analytics embedded in their system is calling it a form of AI. It's confusing the marketplace, and frankly, I think it's incredibly unfair to the consumer."
So what exactly is artificial intelligence? "I like to think of AI ... as being a concept that was initially devised back in the '50s to make computers more useful and capable of independent reasoning," Lovejoy says.
In this interview with Information Security Media Group (see audio link below photo), Lovejoy discusses:
- Deep learning, Bayesian models and unsupervised machine learning versus supervised machine learning;
- The quest for full system autonomy, and what it will take to get there;
- Gaming the system: Why tricking AI or weaponizing AI is difficult;
- Beyond AI marketing messages: Try before you buy.
Lovejoy is CEO of BluVector, which offers a machine learning threat detection and hunting solution. She previously served as head of the business unit at defense contractor Northrop Grumman, from which BluVector was spun out as a stand-alone business in January 2017. Previously, Lovejoy served as president of Acuity Solutions, general manager of IBM's security services division, and CISO of IBM, among other roles.