And, in any case, if anything is a TM it is the brain-body-environment system as a whole (as the world is (part of) the tape).
-
-
Replying to @IrisVanRooij @micahgallen and
I kind of get the argument, but we have the connotative issue. The association with computational does not extend that far, I would guess, but call up the metaphor of brain as a pc/mac/mainframe. What words mean is....fuzzy
#JustTaughtPhilOSci1 reply 0 retweets 0 likes -
Replying to @asehelene @IrisVanRooij and
Not all NNs are Turing complete. So I'm not even sure the people making that argument are aware of what Turing complete even means...? To show a system, e.g., a programming language is Turing complete all you need to do is show it can express NAND. It's really simple.
2 replies 0 retweets 4 likes -
Replying to @o_guest @asehelene and
Mmmm, not quite. You need to show it can implement any arbitrary set of NAND operations. Not so simple, but something that multilayer and RNNs can do.
1 reply 0 retweets 0 likes -
Replying to @tyrell_turing @o_guest and
I thought recurrent connections necessary to get the infinite memory (otherwise not finite number of layers?).
1 reply 0 retweets 2 likes -
Replying to @IrisVanRooij @o_guest and
No, just infinitely wide, potentially.
1 reply 0 retweets 0 likes -
-
Replying to @IrisVanRooij @tyrell_turing and
Recursion seems to me infinitely more plausible than infinite sized networks.
1 reply 0 retweets 3 likes -
-
Replying to @tyrell_turing @IrisVanRooij and
"You need to show it can implement any arbitrary set of NAND operations." For the NN: yes. For a programming language, that's already a given.
1 reply 0 retweets 1 like
So yeah, I should have been clearer.
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.