https://www.nature.com/articles/s41467-019-11786-6 … A very sensible article! If neural nets are to replicate animal intelligence, researchers will have to “decode” a relationship between network architectures and behaviors (the way CNNs are a network representation of translation invariance).
-
Show this thread
-
This article is a good explanation of my “weak AI skeptic” position; I think it’s not plausible that we’ll build a machine that can do everything a mouse can without major innovations in network architecture.
3 replies 0 retweets 10 likesShow this thread -
Replying to @s_r_constantin
When you say 'network architecture', what do you mean? More precisely, how broadly would you define 'network'? I ask as I'm not sure if this is meant as a strongly restricted subset of software or not, & if so then how exactly.
1 reply 0 retweets 0 likes -
Replying to @jprwg
Network architecture = what weights are connected to what other weights, in an ANN. I think neural networks are a strict subset of software, yes.
1 reply 0 retweets 1 like -
Replying to @s_r_constantin
Thanks for the reply. Would you say brains are networks, and/or that AGI will be a network? (And, why do you think so or not so?)
2 replies 0 retweets 0 likes -
Replying to @jprwg
I'm not sure brains are well modeled by networks, no. I think neural structure is suggestive evidence that something network-shaped is has an important function, but I wouldn't be shocked if we discovered that, I dunno, a lot of thinking involved microglia or something.
1 reply 0 retweets 2 likes -
Replying to @s_r_constantin @jprwg
To be clear, it depends on your level of abstraction. Can you map some kind of an ANN to a brain in such a way that nodes correspond to neurons? I'm not certain. Can you "model" a brain with an ANN with some other mapping? Universality results say yes, as I understand.
1 reply 0 retweets 1 like
"Will the first mammal-level AI be an ANN"? That's a social question. Right now the convention is to formulate machine learning programs as ANNs, so most progress is on ANNs. Genetic algorithms, for instance, are an alternate but equivalent formulation:https://openai.com/blog/evolution-strategies/ …
-
-
Replying to @s_r_constantin @jprwg
If the ML field starts using a different formalism in preference to ANNs before we have mammal-level AI, then I guess "the AI won't be a neural net", but in a boring sense. Is the AI going to do some kind of statistical inference, estimating a function based on data? It's gotta.
0 replies 0 retweets 1 likeThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.