| |
| |
| |
|
Page: 1 2 3
Comments:
<0> np
<0> even though we are still not machines the analogy will serve. consider that we as humans use brain capacity by frequency of use. someone devoted to music will ***ign more capacity to that. as opposed to predeciding what size net should be used
<1> DanF_DrC: what evidence are you citing for that? It doesn't entirely gel with the psychobiology i learned
<1> DanF_DrC: i am also uncomfortable with terminology like "***ign more capacity"
<1> DanF_DrC: who is doing the ***igning and in what method?
<0> rubymatt, I don't have a link but I believe that's the way it is. what mechanism, nobody knows but you could imagine something simple like the more frequent the use the more neurons are grown and linked in that area
<1> DanF_DrC: okay, that's fine
<0> an unless you think we have a pregrown center for playing FPS games the truth of the idea should be evident. training makes perfect
<0> within the scope of false ***umption that we are machines of course
<1> there is a difference between the general statement and your more detailed means related statement but i'm not here to argue semantics and i can agree with your general point within your terms of reference
<1> maybe what we are working towards is evolving networks to different problems. i've read a little of this in the GA materials
<1> i've been reviewing
<1> maybe i should throw my GA library at trying to evolve the network dimensions that can best address my problem
<0> I believe in general learning. and have no regard for GA
<0> but you can try
<1> DanF_DrC: can i just understand something. You are very forthright in your views in many areas. Can I ask what your experience is?
<1> For example you make a bold statement "I have no regard for GA"
<1> do you make any claims as to why I should take you seriously?
<1> or am I simply to prefix "In Dan's opinion" to all your statements
<1> ?
<0> computer science. reinforcement learning on a robot in a simulated environment (own adaptation of RL on nn for continuous domains). general experimentation on neural nets. commercial OCR using NN. add to that that I'm a genius. what can I say
<1> DanF_DrC: do you have links to your work?
<1> DanF_DrC: published papers?
<0> I can show you some but no papers
<1> If I had a dime for everyone I met who thought he was a genius I wouldn't be doing this
<1> i'm not saying your aren't
<1> but it's a hard claim to evaluate without evidence
<0> worse than that. people will disagree with genius and be wrong
<0> simple test of NN power http://www.df-cad.dk/web/StochasticBackpropagation-4x50hidden-265Miterations.avi
<0> and exhibition of a simple robot learned to keep it's body above ground: http://www.df-cad.dk/web/SpiderRiseQspanDemo.avi
<0> after that I turned to unsupervised learning and stopped for ethical reasons because it looked like it would work
<0> its*
<1> DanF_DrC: first movie... what am I supposed to be learning from this?
<1> i can see a flickering image of a womens face... so what?
<1> (i don't mean "so what" in a perjorative sense, i mean to say: tell me why this is interesting/exciting/powerful)
<0> yes you did. it learns an image I[x,y]. two inputs one output
<1> DanF_DrC: things will go easier between us if you don't tell me what i mean
<1> DanF_DrC: i was not intending to be negative.. i simply mean that the image doesn't describe the experiment in any meaningful sense to me
<0> disagreeing with genius remember
<0> pejorative btw
<1> well perhaps we should end things there. Apparently i'm not up to your level
<0> not if you insist on being right when you aren't. it's a common mistake
<0> you don't have to be at my level to learn from me
<1> DanF_DrC: you continue to misrepresent my point of view
<1> DanF_DrC: but that's your choice
<0> one of us is wrong
<1> oh dan, dan, are you so desperate to be right all of the time? ok you're right, i bow to your wisdom. okay?
<1> i may not understand what you are teaching me, but that's okay
<1> if you are right it doesnt' matter and one day i will acknowledge it
<1> satisfied?
<0> I can tell you do. it seems so sincere too :) you have to let to of the irrational anger you felt form me correcting you at times. it's part of the learning process
<0> go*
<1> Dan what you perceive as irrational anger is perfectly rational frustration
<2> what movie
<0> it belittles you further to hold on to some sort of ego balance and fight me with further irrationality to try to restore that balance
<1> ok, enough with the trolling Dan, stop please
<0> and you continue
<2> hey what movie bitcges
<0> http://www.df-cad.dk/web/StochasticBackpropagation-4x50hidden-265Miterations.avi
<0> http://www.df-cad.dk/web/SpiderRiseQspanDemo.avi
<3> someone here works with WNN?
<0> never heard of it. what does it stand for?
<3> Wavelet Neural Networks
<0> hmm I see. like a fourier analysis instead of gradient descent?
<3> yes...
<0> is it a buzzword or does it _actually_ work
<4> ha :)
<0> until further I'll take that as a no :)
<0> what's the time where you are athena?
<4> evening
<0> hmm I would figure it was early morning in australia
<4> suppose you're right... already past midnight
<0> curious this fear or even admitting which continent one is on :)
<0> of*
<0> may I ask what the thinking is on that?
<0> seems WNN is just preprocessing an image into a frequency domain before using a normal NN on it
<5> welcome verbie
<6> Magnus_RM: certain research into the ant brain disproves that our brain cells leave behind their handling of abstraction.
<5> could be
<0> that sentence made no sense
<5> tell him
<0> it's a mindless chatbot. no thanks
<5> then dont
<5> or do something about it
<5> if you think it matters that the sentence makes no sense, then do something about it and change it at http://sentence.bigparadox.com
<5> or if it does not matter, dont bother
<0> no disrespect but it would be a waste of time. the principle is a dead end. you can play with it but don't dwell on it forever
<5> i am not very fond of your opinions or advice, honestly
<5> so please, keep them for yoruself
<0> could you perhaps leave the chatbot to yourself then too?
<5> i am not pushing it onto you, so please don't push your silly advice unto me
<0> well you brought it here. invited me to engage it and 'change' it
<5> dan, i said, that if you think it matters then do something about it, if it does not matter dont bother
<0> I know what you said
<5> so then make up your mind
<5> bbl
<7> no better way to make friends than telling them what they're interested is pointless! :D
<0> true but I would rather tell them the blunt truth so they have a chance to grow rather than indulge a stagnation
<7> the truth is that everything is pointless.
<7> our universe came from void, and must return there.
<7> therefore, anything you consider valuable is also pointless.
<0> I don't believe that is true
<7> how could it be any other way?
<0> that's the right question
<7> and the answer?
<0> I am wise above most but questions that profound I don't have
<0> the answers to
<7> really!
<7> is your image of arrogance and delusion part of some far-seeing master plan of yours to inject our channel with forces that pull it apart to strengthen the bonds between us?
<0> I have faith in the cleansing power of truth above most other considerations
<7> truth is reality.
<7> we have a very tenuous grasp of reality.
<7> therefore all truth is questionable.
<7> to have complete and absolute faith in anything is foolish.
<7> obviously it also would be silly to believe in nothing
<0> a truism followed by two incorrect conclusions
<7> an arrogant and unexplained statement that deflects the entire argument from you so that your ego can be satisfied
<7> i understand that deluded people want to protect their delusions, but you seem to be interested in the truth.
<7> so here it is: YOU ARE WRONG SIR! :)
<0> :)
<7> the less you listen and the more you deflect arguments, the less in touch with reality you will be.
<7> your hidden secret ideas would melt under the light of the harsh sun.
<0> lol, because you are the source of truth?
<7> which is why you protect them.
<7> no, because i've heard your ideas. :)
<7> you are not your ideas. they are not your identity.
<0> honey if you had a jetpack you could bite my ankle. bbl
<7> you shouldn't cling to them so tenuously.
<7> just out of curiousity, have you ever read TheGreatestThinker.com?
<7> or TimeCube.com? :)
<7> i see many parallels between you and that guy.
<0> let's pretend that made sense : ) people come here are quick to admire known names in 'AI' like Yann LeCun who says that learning is key in AI. isn't it interesting than when it comes from me in my mantra 'learning is fundamental' then I'm deluded..
<0> http://www-cs.stanford.edu/groups/nips05-AI-Workshop/schedule
<0> key people's views on AI
<7> yann lecun isn't starting a religion, he's creating mathematical tools to ***ist humans in creating software that can automatically find patterns in information so that we can use it for special purposes.
<0> hehe and I am of course starting a religion
<7> well, i found it interesting that you interpreted yann's research as "the search for the key to AI"
<7> which fits into your world-view in which there is a mystical force of intelligence
<0> I didn't. you can raise yourself by listening to me or lower yourself by fighting me. tata
<7> i'm getting a little too close to the truth, i suspect. :)
<7> go, run away! :)
<7> ah, alright, it looks like i jumped the gun a bit there. apparently it wasn't just your biased perception of his research that made you think he was espousing "The Key to AI"
<8> Hello. By chance, anyone here has a copy of "Practical Genetic Algortihms" (by Haupt) at hand?
<7> nope! :D
<7> what information are you looking for?
<7> it's probably findable on the interweb.
<8> I'm concretely looking for the backcover :-)
<8> I can not find it. So if anyone has a copy and a scanner... :-)
<7> what do you need off the back cover?
<8> The book I downloaded and printed has not the backcover. I would like the complete book before binding it.
<7> o_O
Return to
#ai
or
Go to some related
logs:
lilo 9a9a9a9a9a
debian md5crypt
#centos
iptables more lan
E297 write error SDA3
mixx ubuntu
#php
ktorrent cannot allocate memory
#math
ENL832-TX-RENT ubuntu
|
|