| |
| |
| |
|
Comments:
<0> Looking for a good programming language... Suggestions?
<1> fit your tools to the task... what do you want to do?
<0> Long term, a program that can propose hypotheses, and test them.
<0> Vague? I know...
<0> Working on that.
<1> yes, that's rather high-level... would need to know more about what you do closer to the code level
<0> Right now I'm just learning programming languages.
<0> Started on forth today.
<1> on a large project you won't just use one language... you need some easy to use interpreted language to automate simple things for proofs of concept, code generation, and so on
<1> why did you choose forth?
<0> athena: Good benchmarks; concise, readable syntax.
<0> I'm also interested in ocaml.
<0> Same reasons.
<2> delphi
<3> Anyone have a good example of getting AI to generalize?
<1> kaeles, what do you mean by generalize?
<1> never mind generalising, i want to know if anyone has AI! :)
<3> like
<3> uh
<3> i dunno anymore
<3> lol
<3> well, basically generalizing a concept or pattern
<3> bleh
<3> i dunno, time for work anyways
<3> :P
<3> thanks
<4> Kaeles: well, if I took in what you want, there are ontologies that are good to deal with concepts, their relations. There are also some methods/techniques to do pattern recognition
<4> data mining is a good field to extract knowledge from data
<4> data is your input and you try to get some kind of knowledge from the data
<1> ows: ontologies are formed by generalising... i think kaeles was asking for examples of that process
<4> data mining
<4> web/text mining
<1> how is generalisation used in those tasks?
<1> topic extraction maybe
<4> to create ontologies, usually are used Natural Language Processing (NLP) techniques
<4> semantic analysis of documents
<4> corpus
<1> yes, but that isn't what i was talking about
<4> you find patterns of relations between words (that can be concepts) in those docs
<4> so, what do you mean
<4> try to be more explicit
<1> what does is mean to "generalise"?
<1> it, not is
<1> once we've got that established, we can address kaeles' question :)
<4> well, it's being hard to really take in what do you want to know from the concept generalise
<4> you can use bayesian inference to generalise
<1> i don't think you understand the concept well if you can't explain it
<4> "draw from specific cases for more general cases"
<4> extract some knowledge from specific data
<4> and generalise to more general data?
<1> the difficult part in generalising is determining the salient features
<1> that's the basis on which we judge whether it was done "correctly" or not
<1> or even being able to recognize features
<1> anyway... lots of examples in purported AI of generalising... but usually the task isn't difficult... few features, easy to extract... simply find the features shared by the most instances
<4> data preprocessing -> feature selection?
<1> what does "->" mean?
<4> has algos like
<1> if you're asking whether feature selection is a type of preprocessing... i suppose you could cl***ify it as such... depends!
<4> <1> the difficult part in generalising is determining the salient features
<4> so I said that you could do data preprocessing using feature selection algos
<4> I'm not asking :P
<1> ok... i'm finding this conversation a bit difficult
<4> well that ? was like if does data preprocessing -> feature selection matches what you want?
<1> bbl
<4> ok
<4> cya :)
<5> hey everyone
<4> hi
<6> hey where can i get a good directed evolution paper
<7> hmm. what kind of directed evolution paper?
<8> hi
<2> hi
<8> Heya doin Dan?
<2> fine thanks :)
<8> any progress with creating an ai?
<3> anyone read anything about getting depth from a single image?
<8> DanF_DrC go skateboard riding.
<8> Thou Shall Not Make The Image Of Man.
<2> JohnnyL, I don't work on it
Return to
#ai
or
Go to some related
logs:
gentoo deroff
syslog-ng catch all filter
#php
jbalcomb
Intel Corporation 82801EB/ER (ICH5/ICH5R) AC'97 make.conf
#php
ubuntu mythtv packager
Xair howto
#php
#asm
|
|