studiot Posted October 2 Posted October 2 (edited) https://www.bbc.co.uk/news/articles/c0lw0nxw71po Now for the first time scientists researching the brain of a fly have identified the position, shape and connections of every single one of its 130,000 cells and 50 million connections. Edited October 2 by studiot 1
dimreepr Posted October 5 Posted October 5 What struck me as I read this, imagine how much more powerful AI could become if we copy n paste that wiring diagram into a supercomputer?
Luc Turpin Posted October 5 Posted October 5 (edited) We still do not know how 130,000 cells and 50 million connections enable a fly to interact with each other and the world around it. Also, brains do not function like computers. Edited October 5 by Luc Turpin
exchemist Posted October 5 Posted October 5 1 hour ago, Luc Turpin said: We still do not know how 130,000 cells and 50 million connections enable a fly to interact with each other and the world around it. Also, brains do not function like computers. No one is claiming we do. Mapping the connections is the first step in learning how it works. That's all.
MigL Posted October 5 Posted October 5 3 hours ago, Luc Turpin said: Also, brains do not function like computers. It's not the small number of cells involved, but the 50 Million connections. Our best supercomputers are currently massively paralleled simple compute engines. The difficult part is the programming that executes on these simple computers for one or many processes to take advantage of the massive parallelism efficiently. Nature and evolution have been 'working' on the brain of insects like the fly, for about 480 Million years. It won't happen anytime soon, but give our guys a little time to figure out true AI.
geordief Posted October 5 Posted October 5 2 hours ago, exchemist said: No one is claiming we do. Mapping the connections is the first step in learning how it works. That's all. I understand that the "learning system" is made up of pairs of neurons. Do these neurons need to be located in proximity to each other or ste there pairs with "ends" at the opposite side of the brain? I wonder is the next step to digitise this representation and to stimulate it with diffferent inputs. I remember that film where Jeff Goldblum turns into a fly https://en.m.wikipedia.org/wiki/The_Fly_(1986_film)
exchemist Posted October 5 Posted October 5 34 minutes ago, geordief said: I understand that the "learning system" is made up of pairs of neurons. Do these neurons need to be located in proximity to each other or ste there pairs with "ends" at the opposite side of the brain? I wonder is the next step to digitise this representation and to stimulate it with diffferent inputs. I remember that film where Jeff Goldblum turns into a fly https://en.m.wikipedia.org/wiki/The_Fly_(1986_film) Haha I was just thinking about that film. But I haven’t looked into how these connections function in the brain. It’s a long way from my stamping ground.
Recommended Posts