in Link Post

Machine Learning’s ‘Amazing’ Ability to Predict Chaos


Natalie Wolchover:

In a series of results reported in the journals Physical Review Letters and Chaos, scientists have used machine learning — the same computational technique behind recent successes in artificial intelligence — to predict the future evolution of chaotic systems out to stunningly distant horizons. The approach is being lauded by outside experts as groundbreaking and likely to find wide application.

There are a couple of disclaimers here, I think. The main one is that this is a case of “for eight times longer than existing methods”, rather than the “for the lifetime of the system” which the title seems to imply. But still: amazing result. Being British, obviously the must exciting potential application for this lies in improved weather prediction. But I’m sure there are other uses, too…

As always, I’d love to know how much computing power this took, but I couldn’t fin that information in either this article or the original paper.