Isn't that the crux of the argument? That is that DRL is similarly scalable as DL and therefore there are no more obstacles other than compute? I believe there is a conceptual obstacle, but I don't think it's a big hurdle!
-
-
Replying to @IntuitMachine @Plinz
There is a good chance we're witnessing a second incarnation of the Church-Turing thesis. My main concern is unsupervised learning: would extra compute solve that particular problem? I don't know, but from the outset it would appear as a primarily conceptual obstacle.
2 replies 0 retweets 0 likes -
Replying to @FieryPhoenix7 @Plinz
This isn't a problem that relates to undecidability. This is learning what is knowable and the problem is that we don't know an efficient algorithm for learning. This brute force can potentially discover this algorithm.https://medium.com/intuitionmachine/the-explosive-ramifications-of-combining-intuition-with-logic-in-deep-learning-610f6f3477da …
1 reply 0 retweets 0 likes -
Replying to @IntuitMachine @FieryPhoenix7
Undecidability is a big red herring caused by the timeless definition of truth semantics in classical mathematics. In a fully computational perspective, it resolves to incompressibility: math attempts a lossless compression to axioms, but ignores realizability of the compression
1 reply 0 retweets 2 likes -
Replying to @Plinz @FieryPhoenix7
I am indeed curious as to why so many researchers latch on to undecidability as a problem that relates to general intelligence? Perhaps you may have a good explanation.
3 replies 1 retweet 1 like -
From the perspective of general intelligence, being able to make good prediction up to a short time horizon is good enough. So the halting problem is no where near being an issue.
1 reply 0 retweets 1 like -
Replying to @IntuitMachine @FieryPhoenix7
I think the halting problem cannot even be formulated in a fully computable paradigm. It is similar to the question if an omnipotent God can create a rock that he cannot lift. Omnipotence may be as ill-defined as infinity.
2 replies 0 retweets 1 like -
Replying to @Plinz @IntuitMachine
FWIW, the halting problem is probably irrelevant to the question of general intelligence. The only reason it exists is to show that there are fundamental limitations to computation, and we're pretty sure general intelligence is a decidable problem.
1 reply 0 retweets 1 like -
Replying to @FieryPhoenix7 @IntuitMachine
I think the halting problem is more profound: it shows that there are fundamental problems with mathematics, which computation resolves (to the degree to which they can be resolved).
1 reply 0 retweets 0 likes -
Replying to @Plinz @IntuitMachine
Do you mean that mathematics itself has inherent flaws or is it just that our understanding of it is fundamentally rudimentary?
1 reply 0 retweets 0 likes
IMHO Gödel has shown that the nonconstructive parts of math are flawed.
-
-
Replying to @Plinz @FieryPhoenix7
That's a very strong statement. Do you have references to this?
1 reply 0 retweets 0 likes -
Replying to @IntuitMachine @FieryPhoenix7
This is actually what it says. Hilbert asked to fix the flaws introduced by infinities as apparent in Cantor's set theory, and Gödel and Turing responded by demonstrating that they cannot be fixed.
1 reply 0 retweets 3 likes - 3 more replies
New conversation -
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.