It would be very convenient (though mind-numbingly boring) if the content of our days was just a simple interpolation of things we had experienced hundreds of times before. We could train a deep learning model to replace us!
Show this thread
-
In fact, every single of our days is completely unique -- unlike any day we've experienced before, and unlike any day experienced by anyone since the dawn of humanity. And somehow we can still cope. That's the point of general intelligence: extreme generalization.
Show this thread
Machine learning is about fitting to a static distribution -- human learning is about gathering knowledge that may turn out to be useful in a future that is guaranteed to share little commonality with the past.
-
-
How do we start looking for the science of AI rather than worrying about beating the SOTA of the engineering models of ML/DL?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
There are lots of ways of doing ML but one of the aims is to make generalisations
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Human learning is about fitting a static distribution to a dynamic instance to minimize the effort spent on commonalities and instead concentrate on the small manageable deviations. ML, as it stands, is just starting on the deviation part.
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.