Very young humans are able to use their limited experience and priors to appropriately respond to situations they've never seen before, and even to situations none of their genetic ancestors have seen before (modern environments). That's "developer-aware extreme generalization"https://twitter.com/rodneyabrooks/status/1271828807305986048 …
-
-
Preach!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Must be probabilistic rules derived from training data and updated by changing results and motivations, right?
-
training sets are the lazy option. eg automatic control essentially tries to uncover the 'secret formulas' to motion. but in many cases solving a problem the proper way is super hard, hence 'training sets'. a solution that's not good enough, however, is as good as no solution
End of conversation
New conversation -
-
-
I would even argue that intelligence is (at least partially) defined by the ability to converge to a correct solution while minimizing information processed. Brute forcing is pretty stupid in a lot of cases.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
since that capacity is really supernatural.. we can only grasp a fraction of how really works..
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Arguably, the development of the brain stem from conception is largely mapped out by DNA and not adaptable. To have axons as long in some neurons you need this weaving dance to progress and wrap thalamus with cortex. Within that architecture adaption happens. DNA is bruteforcish
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
It helps to have heritable architectural priors tuned via a genetic algorithm (evolution). My keras models start with an explicit declaration of a principal layer type that I have LEARNED to choose based on fundamental attributes of the problem domain. This mapping is learnable.
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.