News for hackers. Stuff that matters.
Re:Tensorflow Deepmath: Experiments towards neural network t... by runesoerensen on Monday November 07, 2016 @06:28PM (#12896240)
Relevant HN discussion about the paper: https://news.ycombinator.com/item?id=12031253
Re:Tensorflow Deepmath: Experiments towards neural network t... by cmrx64 on Monday November 07, 2016 @09:06PM (#12897027)
Neat! I look forward to seeing what they come up with :) For similar sorts of things, check out NeSy: http://daselab.cs.wright.edu/nesy/NeSy16/
Re:Tensorflow Deepmath: Experiments towards neural network t... by bitL on Monday November 07, 2016 @07:57PM (#12896721)
What logic does it use? Any chance using relevance or linear logic to avoid vacuous truth proofs (like in set theory)?
Re:Tensorflow Deepmath: Experiments towards neural network t... by cmrx64 on Monday November 07, 2016 @09:05PM (#12897021)
It's currently integrating higher-order logic (HOL) with a reimplementation of HOL Light.
Re:Tensorflow Deepmath: Experiments towards neural network t... by jeff_marshall on Monday November 07, 2016 @09:10PM (#12897042)
I wonder what the reason for the re-implementation is. Perhaps the existing kernel isn't fast enough at evaluation for training (wild guess)? I'll have to RTFM...

(update) their doesn't appear to be any FM. Can anyone post any links to papers/lists/etc.?

Re:Tensorflow Deepmath: Experiments towards neural network t... by danblick on Monday November 07, 2016 @09:58PM (#12897250)
Check out the "Deepmath" paper linked above. (The paper is from a few months ago, the source code is new IIUC.)

https://arxiv.org/abs/1606.04442

Re:Tensorflow Deepmath: Experiments towards neural network t... by jeff_marshall on Monday November 07, 2016 @10:45PM (#12897446)
thanks!

(edit) I still don't see any rationale for the theorem prover kernel re-implementation. Again, pointers appreciated.

Re:Tensorflow Deepmath: Experiments towards neural network t... by dgacmu on Tuesday November 08, 2016 @08:59AM (#12900233)
HOL is written in OCaml. The likely answer is because integrating OCaml and TensorFlow would be painful -- and OCaml isn't one of the languages Google supports internally [1]. Probably was easier to reimplement than to try to do all the contortions needed to use it in this context.

(I work part time on the Brain team, but I haven't asked the authors about this -- this is a best guess based upon knowing TF, not an informed statement.)

[1] https://www.quora.com/Which-programming-languages-does-Googl...

Re:Tensorflow Deepmath: Experiments towards neural network t... by posterboy on Tuesday November 08, 2016 @02:49AM (#12898421)
My bet is on concurrency optimizations
Re:Tensorflow Deepmath: Experiments towards neural network t... by Mao_Zedang on Monday November 07, 2016 @09:19PM (#12897073)
Hook it up to a network that makes theorems.
Re:Tensorflow Deepmath: Experiments towards neural network t... by SticksAndBreaks on Tuesday November 08, 2016 @05:21AM (#12899041)
Every problem i can break into a sub problem is solveable. Thus proofen by induction that i can break recursively down every problem, i deduce- that i crashed with a NN-Stackoverflow.
Re:Tensorflow Deepmath: Experiments towards neural network t... by thomasahle on Tuesday November 08, 2016 @07:14AM (#12899597)
> Every problem i can break into a sub problem is solveable.

Given the sub problems are either solvable or can in turn be broken down, I suppose?