Probabilistic programming does in 50 lines of code what used to take thousands

#1

Post by ub » Mon, Apr 13, 2015, 4:09 PM UTC

Probabilistic programming does in 50 lines of code what used to take thousands
newsoffice.mit.edu

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#2

Post by newpattern » Mon, Apr 13, 2015, 4:19 PM UTC

Anybody here on HN have experience with probabilistic-programming ? This looks quite disruptive if it works.

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#3

Post by ub » Mon, Apr 13, 2015, 4:21 PM UTC

No experience but looks like they are organizing a summer school on probabilistic programming languages. http://ppaml.galois.com/wiki/wiki/SummerSchools/2015/Announc...

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#4

Post by pliny » Mon, Apr 13, 2015, 4:58 PM UTC

relevant paper: http://mrkulk.github.io/www_cvpr15/1999.pdf

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#5

Post by eximius » Mon, Apr 13, 2015, 5:02 PM UTC

newpattern wrote: post #2

Anybody here on HN have experience with probabilistic-programming ? This looks quite disruptive if it works.

Really? "Disruptive"?

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#6

Post by stevenspasbo » Mon, Apr 13, 2015, 5:07 PM UTC

Is that 50 lines of code, or 50 lines of using a library that's thousands of lines of code?

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#7

Post by jb55 » Mon, Apr 13, 2015, 5:08 PM UTC

If anyone wants to jump into this, Josh Tenenbaum and Noah Goodman put together this amazing interactive book for learning probabilistic programming with Church: https://probmods.org/

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#8

Post by aet » Mon, Apr 13, 2015, 5:20 PM UTC

stevenspasbo wrote: post #6

Is that 50 lines of code, or 50 lines of using a library that's thousands of lines of code?

I agree -- I think lines of code is the wrong measure. A better measure would be how large is the set of problems that occur in practice that normally would take thousands of lines which can now solve be solved with 50 lines of code using a probabilistic programming language.

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#9

Post by nanofortnight » Mon, Apr 13, 2015, 5:24 PM UTC

newpattern wrote: post #2

Anybody here on HN have experience with probabilistic-programming ? This looks quite disruptive if it works.

If you want to know about the history of similar language paradigms, look at logic programming. Similar to probabilistic programming, it is declarative, utilises an engine at run time, the user builds a model of the word then queries it, and it uses fewer lines of code to express for problems within a limited domain.

Hardly "distruptive". You should really think of them as DSLs that people actually use.

If you're intrested in such things, perhaps you should take a look at PRISM, which embeds a probabilistic framework within B-Prolog. http://rjida.meijo-u.ac.jp/prism/

Re: Probabilistic programming does in 50 lines of code what used to take thousands

#10

Post by nanofortnight » Mon, Apr 13, 2015, 5:29 PM UTC

aet wrote: post #8

stevenspasbo wrote: post #6

Is that 50 lines of code, or 50 lines of using a library that's thousands of lines of code?

I agree -- I think lines of code is the wrong measure. A better measure would be how large is the set of problems that occur in practice that normally would take thousands of lines which can now solve be solved with 50 lines of code using a probabilistic programming language.

Generally it's anything that you can express using a probabilistic graphical model. It's mainly useful in engineering (real engineering), control systems, robotics, inference, machine learning and artifical intelligence circles.