About promising too much, the next #AI winter, and that even with the nicest & most expensive hammer in your hands the world still doesn’t consist of nails only.
#ml #rl #dl #aiAfterDLhttps://twitter.com/GaryMarcus/status/1161690752524550144 …
-
-
Replying to @SwarmDynamics
Even this year when I talked about AI winters and how the pattern repeat again and again in a super predictable way I had people tell me "no but DNNs are different, it's not hype if you promise the moon as long as you believe you can deliver"
1 reply 0 retweets 2 likes -
Replying to @sina_lana @SwarmDynamics
The difference for me is, even if we stopped with what DNNs could demonstrably and easily do in 2017 and never discovered anything ever again, that's still enough to build an entire industry around. Computer vision isn't going to stop being useful. But could DRL crash? Sure.
1 reply 0 retweets 1 like -
Replying to @ngutten @SwarmDynamics
The other stuff that got hyped before the other winters didn't disappear either ¯\_(ツ)_/¯
1 reply 0 retweets 0 likes -
-
If so,
@GaryMarcus himself,@mpshanahan and a few others are trying to resurrect it :). Probably it never died though? I’m sure a few people kept it going ‘underground’, waiting for the time it wouldn’t be automatically dismissed anymore. 1/22 replies 0 retweets 0 likes -
Replying to @manuelbaltieri @ngutten and
For the record, I'm _not_ trying to resurrect symbolic AI (complete with symbol grounding problem). I'm interested in incorporating certain ideas from symbolic AI into a deep learning framework (eg: relational representations and compositionality). See https://arxiv.org/abs/1905.10307
1 reply 0 retweets 8 likes
I am not per se either; i want to resurrect symbol-manipulation in the context of hybrid systems that also incorporate deep learning.
-
-
Replying to @GaryMarcus @mpshanahan and
Grounding problem, bootstrapping, thinking outside the code.. these are all the same problem. ML doesn't need to address it, but AGI research surely does.
0 replies 0 retweets 0 likesThanks. 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.