One of my favorite related pieces on the subject: http://www.michaelnielsen.org/ddi/if-correlation-doesnt-imply-causation-then-what-does/ …
I'm quite sympathetic to Judea Pearl's point of view that AI needs some notion of causality built in at the roots: https://www.quantamagazine.org/to-build-truly-intelligent-machines-teach-them-cause-and-effect-20180515/ …pic.twitter.com/ttiMJgkFP2
-
-
-
Did you realize I wrote it?
- 2 more replies
New conversation -
-
-
This Tweet is unavailable.
-
I haven't read the whole thing, but I've read a fair chunk, much of it very closely. My post http://www.michaelnielsen.org/ddi/if-correlation-doesnt-imply-causation-then-what-does/ … is based on it.
End of conversation
-
-
-
This Leon Bottou talk outlines his research on finding causality in data with modern ML techniques. Very interesting early results!https://www.youtube.com/watch?v=DfJeaa--xO0 …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The agent-viewpoint (agent, actively controlling its environment to achieve some goals, see Russel-Norvig book, all work on RL/MDPs) is inherently causal. So we can refine the way we do things, but this would not be a fundamental change.
- 5 more replies
New conversation -
-
-
Agreed. It is still remarkable to me that people don't accept probability alone is not enough for causal reasoning. That said I don't think it is as easy as he makes like. All the hard work isn't over, especially in the context of massively complex causal systems.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Deep learning is just curve fitting. https://wp.me/p47Kt7-1uU To build truly intelligent machines, teach them cause and effect.pic.twitter.com/JO9Qjzq5Js
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is so fascinating. Still attempting at his 'Causality' (second edition). Quite worth the effort.pic.twitter.com/4trd7wsTwi
Thanks. 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.