Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How can one manage thousands of IF…THEN…ELSE rules?

programmers.stackexchange.com

Post reply on HN
Print view View original

81–86 of 86 posts

Re: How can one manage thousands of IF…THEN…ELSE rules?

#81

Post by Dn_Ab »

StavrosK wrote: post #80
Dn_Ab wrote: post #77

Earlier quoted context omitted.

Oh cool =) I don't really think in terms of classification and optimization anymore and don't really see much distinction between minimizing a loss or objective function. But I did rant on because I am really passionate about how much unnecessary complexity (hehe) is in much of machine learning. And am always eager to talk about promising unifying approaches. If you haven't read up on semirings or computational algeb…

Oh man, I can't stand math :/ It's one of my weaknesses when it comes to ML, I really have to struggle with the formalism...

Yeah. It must be pretty energy intensive for the brain consider all that rewiring that occurs. I take the interlaced GIF approach to learning hard things.

Re: How can one manage thousands of IF…THEN…ELSE rules?

#82

Post by StavrosK »

Dn_Ab wrote: post #81
StavrosK wrote: post #80

Earlier quoted context omitted.

Oh man, I can't stand math :/ It's one of my weaknesses when it comes to ML, I really have to struggle with the formalism...

Yeah. It must be pretty energy intensive for the brain consider all that rewiring that occurs. I take the interlaced GIF approach to learning hard things.

Top-down?

Re: How can one manage thousands of IF…THEN…ELSE rules?

#83

Post by seanmcdirmid »

teamonkey wrote: post #78
seanmcdirmid wrote: post #58

Earlier quoted context omitted.

Often through physics engines. You express a bunch of constraints then let a physics engine solve them on each time step (and sometimes across multiple time steps). Custom (non-physics) solutions involve declarative sets of constraints being solved by some sort of engine; such solutions often start to resemble physics engines even if the physical rules even if the constraints are not exactly physical. Learning how to…

Using a physics engine to express an AI problem? That's an interesting solution, but that's certainly the first time I've heard of one being used like this?

But you know, the problem as stated was how a bunch of influences affect...cows. So think of them as a bunch of mechanical cows instead, and the influences are modeled through springs and collisions, things like that.

AI is often cast as constraint solving, and a physics engine is basically a constraint solver. Of course, for real the AI engine runs next to the physics engine. Ah! I should have mentioned Rodney Brooks' work:

http://en.wikipedia.org/wiki/Subsumption_architecture

where you model intelligence as a bunch of simple prioritized reflex-like behaviors. Now, on each step, you have each "entity" execute their reflexes according to the current environment, where some reflexes act to disable others. The reflexes are incredibly simple, you probably couldn't even express A* path finding, but you could have the act as ants following a diffusely scented trail. Also see his paper "Elephant's Don't Play Chess."

Re: How can one manage thousands of IF…THEN…ELSE rules?

#84

Post by exDM69 »

cartosys wrote: post #76
exDM69 wrote: post #70

I suggested using Prolog in the comments. His problem description was not quite good enough for me to assess whether or not Prolog will be a good choice, but at least it's a candidate. A Prolog program is essentially a list of rules that are written as an implication. Something like: moves_to(Cow, Location) :- hungry(Cow), current_location(Cow, OldLoc), food_in(OldLoc, OldFood), food_in(Location, Food), Food > OldFoo…

And we're waiting for a link.

There you go: http://programmers.stackexchange.com/a/165126/65434

Re: How can one manage thousands of IF…THEN…ELSE rules?

#85

Post by bumbledraven »

exDM69 wrote: post #70

I suggested using Prolog in the comments. His problem description was not quite good enough for me to assess whether or not Prolog will be a good choice, but at least it's a candidate. A Prolog program is essentially a list of rules that are written as an implication. Something like: moves_to(Cow, Location) :- hungry(Cow), current_location(Cow, OldLoc), food_in(OldLoc, OldFood), food_in(Location, Food), Food > OldFoo…

Good answer. Nice to see the original questioner accepted it. To take it a bit further, so-called "answer set solvers" like clasp (http://potassco.sourceforge.net/) provide a fast way to run these kinds of logic programs.

Re: How can one manage thousands of IF…THEN…ELSE rules?

#86

Post by ramgorur »

Dn_Ab wrote: post #34
SeanLuke wrote: post #13

Earlier quoted context omitted.

All machine learning techniques build models for you. ADTrees, Random Forests, and Decision trees are all classifiers, so they serve this guy's needs. Genetic Programming is not a machine learning technique: it is an optimization method for a very specific task and is quite unsuited for his purposes.

No, I am certain you are confusing Genetic Programming [1] for Genetic Algorithms [2]. I am a fan (do a lot with them) of the prior but have never really been impressed by the latter. Wikipedia says Genetic Programming is a specialization of genetic algorithms where the objects are programs. I think this is the only thing I have ever seen where the specialization turns out to be more general than the parent. I'd also…

"Wikipedia says Genetic Programming is a specialization of genetic algorithms where the objects are programs. I think this is the only thing I have ever seen where the specialization turns out to be more general than the parent."

I think GP is "not only" for evolving assembly programs -- http://www.genetic-programming.org/combined.html

Post reply on HN