"These are not breakthrough ideas. They're not wildly creative, but they seem like pretty cool ideas that somebody might try," admits Jeff Clune, the professor who leads the UBC lab.
As amazing as today's AI programs can be, they are limited by their need to consume human-generated training data. If AI programs can instead learn in an open-ended fashion, by experimenting and exploring 'interesting' ideas, they might unlock capabilities that extend beyond anything humans have shown them.
The project demonstrates an early step toward what might prove a revolutionary trick: letting AI learn by inventing and exploring novel ideas.
Clune's lab had previously developed AI programs designed to learn in this open-ended way, allowing them to identify what's most intriguing without human-defined parameters.
Collection
[
|
...
]