“By Teaching Computers ‘Regret,’ Engineers Hope to Teach Them to See the Future.”

Artificial Intelligence researchers have struggled for decades to create computers that can understand the range of human emotions and feelings, but a team of researchers at Tel Aviv University simply wants to make them feel regret. Working with funding from Google, they hope to make computers understand what it’s like to pursue an outcome only to be disappointed. That, they think, could really help computers predict the future.

While software may never know what it’s like to roll out of bed with splitting headache and dress quietly in the dark, it can certainly measure the distance between a desired outcome and the actual outcome achieved. And by doing so computers could learn to minimize “regret,” which in this case is measured by that distance.

TAU computer scientists working on learning theory and other thorny computer intelligence issues think that by teaching computers to reduce regret, they would essentially be teaching them to evaluate all the relevant variables surrounding an outcome in advance.

– Clay Dillow @ popsci.com

Speaking as a computer programmer, I feel something like regret when my programs don’t work. Conversely, when I fix a program and it works, regret gives way to happiness.

A computer program is a projection of a human mind, or minds plural. How natural, then, to use the metaphor of regret as a programming construct.

I find the idea poetic. Not really a technical breakthrough, perhaps — oh, the math is probably faster or something — but the basic idea, making lots of calculations until the numbers say “Do This” or “Stop Doing This” … well, that’s as old as binary programming. A new name for better math: poetry for programmers.

On the other hand, they’ll use it to seduce you with advertising, monitor your location and behavior, anticipate … every need.

~~ Karl Jones