Machine Learning for Structured Finance

After picking up some passengers, a Las Vegas shuttle bus took off on its maiden voyage. At some point along its route, the shuttle's path was blocked by a truck.1 The shuttle dutifully stopped as the other vehicle was trying to back up into an alley. As the truck backed up closer to the shuttle, the latter did not move or honk its horn. The shuttle's passengers watched in astonishment as the truck and the shuttle got into a slow-motion fender bender. Fortunately, no one was hurt. Why did the shuttle's driver not just move back or sound the horn? Even a green driver knows that if a semitruck is backing into you, you move. You do not need a driver's license to know that.This shuttle, however, did not have a driver. The shuttle was self-driving-on its first drive around downtown Las Vegas.2 The incident offers an illustration of some of the challenges of artificial intelligence (AI) and machine learning. Although these new technologies are game changers, like other historical technological revolutions they have downsides and weaknesses. This article attempts to describe machine learning in this perspective. © 2017 Institutional Investor LLC.