- Joined
- May 18, 1997
- Messages
- 55,626
This is why we can't have nice robots. You build a bunch of smart robots, and what do they do? They build smarter robots. And you know what those robots do? Right, they build even smarter robots. And you know what those robots do? Those bastards sneak up on you in the middle of the night....because they were smart enough to build in night-vision...and those unholy monstrosities gut you like a 12-point buck in back woods of Tennessee, but probably use another robot instead of a tree to hang you up in. But all that really does not have anything to do with this, or does it?
The Google researchers created a machine learning system that used reinforcement learning—the trial and error approach at the heart of many of Google’s most notable AI exploits—to figure out the best architectures to solve language and image recognition tasks.
Not only did the results rival or beat the performance of the best human-designed architectures, but the system made some unconventional choices that researchers had previously considered inappropriate for those kinds of tasks.
The Google researchers created a machine learning system that used reinforcement learning—the trial and error approach at the heart of many of Google’s most notable AI exploits—to figure out the best architectures to solve language and image recognition tasks.
Not only did the results rival or beat the performance of the best human-designed architectures, but the system made some unconventional choices that researchers had previously considered inappropriate for those kinds of tasks.