aviatoreb Posted September 3 Report Share Posted September 3 3 minutes ago, EricJ said: Absolutely. I was attempting to make a double joke, as "Reviewer 2" is a subject of many jokes and memes for doing horrible things (mostly incompetence) during the anonymous review process. I caught your joke - Im sorry, I didnt mean make it seemed otherwise. 1 Quote Link to comment Share on other sites More sharing options...
Scott Dennstaedt, PhD Posted September 5 Report Share Posted September 5 Who remembers neural networks? It wasn’t all that long ago that the topic of neural networks was front and center in the world of weather forecasting. That soon ran into a dead end given the computing resources at the time were not sufficient to solve these difficult problems. Therefore, weather forecasting and AI have a substantial hill to climb. 1 Quote Link to comment Share on other sites More sharing options...
Vance Harral Posted September 5 Report Share Posted September 5 2 hours ago, Scott Dennstaedt, PhD said: Who remembers neural networks? It wasn’t all that long ago that the topic of neural networks was front and center in the world of weather forecasting. That soon ran into a dead end given the computing resources at the time were not sufficient to solve these difficult problems. Modern AI is almost entirely based on those same neural network principles. It's been mostly rebranded as "deep learning", but the "deep" in that jargon phrase doesn't indicate anything philosophical. It's just that the number of ranks in the neural network (i.e. it's depth) can be significantly larger, due to the availability of dramatically higher compute power. In particular, the simple-but-highly-parallel vector processors developed for high-end computer graphics in the late 90s and early 2000s turn out to be really good at neural net calculations. In 2006, Nvidia rolled out a programming library called CUDA to facilitate doing this on their GPU chips, and they've been an AI darling ever since. I'm not smart enough to know what effect all that has on modern weather forecast modeling, so I'll refrain from commenting on that. But whatever dead end neural networks ran into back in the 90s have long since been plowed wide open. Quote Link to comment Share on other sites More sharing options...
Scott Dennstaedt, PhD Posted September 5 Report Share Posted September 5 Yes, two decades ago, the computing power was the limiting factor. However, with the emerging technology of quantum computing, this is now on a more level playing field. I am writing a piece for FLYING magazine that should be published next year that will discuss this topic a bit more with regard to weather forecasting. 2 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.