Improving Business Decision Making with Bayesian Artificial Intelligence
GOTO
bayesian
artificial intelligence
decision making
keynote
GOTO 2017 talk on why current AI falls short for business decisions and how Bayesian methods bridge perception and inference.
GOTO 2017 - Copenhagen, Denmark
In this talk I argue that while we’ve made progress in AI and machine learning, we’re still playing around in the paddling pool. Perception (recognizing images, parsing text) is only half the story. The other half is inference: reasoning under uncertainty, combining prior knowledge with data, and making decisions that honestly account for what we don’t know.
I walk through the landscape of AI and machine learning, explain why current approaches fall short for business decision making, and make the case for Bayesian methods as the principled way to bridge the gap between perception and inference. The talk also touches on how our brains reason probabilistically, and why our models should too.