How to Build Artificial Intelligence We Can Trust
Computer systems need to understand time, space and causality. Right now they don’t. by @GaryMarcus and Ernest Davis
https://www.nytimes.com/2019/09/06/opinion/ai-explainability.html …
#AI #MachineLearning #2MA
-
Causality is gonna be the hard one IMO. Inferring causality at the appropriate level of abstraction is something people even suck at - I know I do.
people aren’t perfect at it, and tend to overattrbute correlation, but they do for example understand objects in terms of their causal functions. when AI can do the same it will be a revolution.
-
-
I just finished Kahneman's "Thinking, fast and slow" and dove straight into hofstadter's "I am a strange loop". If you know any of those, you'll see why I felt compelled to chip in
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
That sounds like a great thing for a computer to be able to model. But how do you train a neural network in this regard? What kind of data-set would you feed it?
It sounds *way more abstract* than, say, keeping a car on the road.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.