Cartography of the space of information, specifically. Which is a Riemannian manifold, rather than a Euclidian space, unlike what we are used to when it comes to actual maps.
-
-
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Nice thread, could we call a map a two dimensional model?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The only difference between pattern recognition&model building is the granularity level of it's Operation: fine in former,coarse in latter. Keyword recognition in compilers fall into the former category while parsing of complete statements (including Keywords) - into the latter.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Aside: As a meandering undergrad I used to be a big fan of work (which unfortunately was never actionable) of cognitive maps and cognitive collages. A somewhat philosophical perspective on space maps http://www.cognitivemap.net/HCMpdf/Ch1.pdf
-
That PDF is the first chapter of this ancient classic http://www.cognitivemap.net/HCMpdf/HCMComplete.pdf …
End of conversation
New conversation -
-
-
Relevant quotes/aphorisms: "All models are wrong, but some of them are useful." - Box "The most accurate map possible would be the territory, and thus would be perfectly accurate and perfectly useless." - Gaimanhttps://twitter.com/fchollet/status/1036142507199881216 …
-
1 Abstractions are not wrong *per se*. 2 Intelligence is also about choosing the right level of abstraction for a representation (which may vary widely depending on the purpose).
End of conversation
New conversation -
-
-
But isnt pattern recognition a map as well? It's a mapping between a noisy observation space and abstract labels.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
How would you model systemic racism in banking? In law enforcement? Class mobility? Those would be interesting models.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
So a map of the world will basically be a complicated decision tree (if else statements). Do you suggest that such a map be prefed or do we allow the machines to create such a map by themselves (via reinforcement learning etc.)?
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.