Avoiding fusion plasma tearing instability with deep reinforcement learningwww.nature.comexternal-link OmnipotentEntity ( @OmnipotentEntity@beehaw.org ) Science • 4 months ago message-square1arrow-up122
arrow-up122external-linkAvoiding fusion plasma tearing instability with deep reinforcement learningwww.nature.com OmnipotentEntity ( @OmnipotentEntity@beehaw.org ) Science • 4 months ago message-square1
minus-square jarfil ( @jarfil@beehaw.org ) linkfedilink7•4 months agoAs some sources have titled it: “AI fixes fusion!” /s On a more serious note, I find it interesting that a 25ms control loop is enough, and that: the tearing prediction model could forecast the instability 300 ms before the disruption Good model.
As some sources have titled it: “AI fixes fusion!” /s
On a more serious note, I find it interesting that a 25ms control loop is enough, and that:
Good model.