Gotos,for loops,foreach loops and list comprehensions are all control flow concepts that increasingly ease programming (note: interventions are akin to setting variables and watching effect on flow). I believe the discrete probability monad ought be next iteration on control flow
-
Show this thread
-
I think calling it a probability monad & the focus on bayesianism is a distraction. Better seen as an evolution of list comprehensions: lazy stream;automatically managing backtracking, constraints and look-ahead search. Such a construct also eases writing of many AI algorithms
1 reply 1 retweet 2 likesShow this thread -
Currently, to most, probabilistic programming = hamiltonian monte carlo. To the rest majority, graphical models. There's too, idea that continuous is always better. This isn't true as it depends on the data structure (fixed vectors? or lazily grown trees?) you're distributed over
1 reply 1 retweet 1 likeShow this thread
I believe Haskell monads are a funny quirk of history. Value-producing delimited continuations are the most general construct underlying all of these patterns. Better to express them directly through functional-logic programming with failure and backtracking (not call/cc).
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.