
Honesty is key to machine learning’s future – Roberts
Oxford-Man Institute director on why tomorrow’s models will gracefully admit defeat

Stephen Roberts likens the habit among some investors of seeing order in data where none exists to ancient astrologers spotting lions, bears and deities among the stars.
As director of the Oxford-Man Institute of Quantitative Finance, Roberts’ research focuses on machine learning algorithms that discern patterns in markets. The trouble is that algorithms, he says, are even more capable than humans of finding “weird and wonderful things” in the numbers.
This tendency to overfit to the data has