--- Log opened Mon Jun 13 00:00:07 2011
00:01 -!- augur [~augur@208.58.6.161] has joined ##hplusroadmap
02:12 -!- alystair [alystair@24-246-14-18.cable.teksavvy.com] has quit [Ping timeout: 240 seconds]
02:26 -!- augur [~augur@208.58.6.161] has quit [Read error: Connection reset by peer]
02:26 -!- augur [~augur@208.58.6.161] has joined ##hplusroadmap
02:35 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has quit [Read error: No route to host]
02:35 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has joined ##hplusroadmap
02:38 < dbolser> can anyone point me at that video hosting site with good subtitle editing functionality?
02:38 < dbolser> I was linked to a vid on that site in here about 6 months ago
02:39 < Utopiah> http://www.universalsubtitles.org ?
02:40 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has quit [Read error: Connection reset by peer]
02:41 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has joined ##hplusroadmap
02:46 < dbolser> Utopiah: your fast!
02:47 < dbolser> thanks
02:48 < dbolser> looks right, which is all I need :-D
02:53 -!- archels [~neuralnet@541ED8B0.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 240 seconds]
03:05 -!- archels [~neuralnet@541ED8B0.cm-5-7d.dynamic.ziggo.nl] has joined ##hplusroadmap
03:20 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has quit [Read error: Connection reset by peer]
03:21 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has joined ##hplusroadmap
03:52 -!- lumos [~lumos@92.20.112.33] has joined ##hplusroadmap
04:14 -!- klafka [~textual@cpe-69-205-70-55.rochester.res.rr.com] has joined ##hplusroadmap
04:24 -!- uniqanomaly [~ua@dynamic-78-8-90-161.ssp.dialog.net.pl] has joined ##hplusroadmap
04:34 -!- lumos [~lumos@92.20.112.33] has quit [Ping timeout: 240 seconds]
04:43 -!- lumos [~lumos@92.20.112.33] has joined ##hplusroadmap
06:12 -!- klafka [~textual@cpe-69-205-70-55.rochester.res.rr.com] has quit [Quit: Computer has gone to sleep.]
06:34 -!- QuantumG [~qg@rtfm.insomnia.org] has quit [Ping timeout: 240 seconds]
09:17 -!- eridu [~eridu@gateway/tor-sasl/eridu] has joined ##hplusroadmap
10:01 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has quit [Ping timeout: 252 seconds]
10:05 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has joined ##hplusroadmap
12:19 -!- uniqanomaly [~ua@dynamic-78-8-90-161.ssp.dialog.net.pl] has quit [Quit: uniqanomaly]
12:46 -!- eudoxia [~eudoxia@r190-133-10-21.dialup.adsl.anteldata.net.uy] has joined ##hplusroadmap
12:58 -!- eudoxia [~eudoxia@r190-133-10-21.dialup.adsl.anteldata.net.uy] has quit [Quit: Leaving]
13:10 < kanzure> senate.gov too? http://lulzsecurity.com/releases/senate.gov.txt
14:34 -!- archels [~neuralnet@541ED8B0.cm-5-7d.dynamic.ziggo.nl] has quit [Ping timeout: 240 seconds]
14:45 -!- archels [~neuralnet@541ED8B0.cm-5-7d.dynamic.ziggo.nl] has joined ##hplusroadmap
14:47 -!- nsh_ [~nsh@cpc3-broo4-0-0-cust997.14-2.cable.virginmedia.com] has joined ##hplusroadmap
14:52 < augur> kanzure: ping
14:54 -!- QuantumG [~qg@rtfm.insomnia.org] has joined ##hplusroadmap
15:03 < nsh_> what's new?
15:11 -!- lumos [~lumos@92.20.112.33] has quit [Ping timeout: 255 seconds]
15:15 < kanzure> augur: pong
15:16 < kanzure> nsh_: pong
15:16 < augur> kanzure: is there a working group here for AI?
15:17 < kanzure> here? what?
15:18 < kanzure> i think that ai is possible and that everyone working on it basically sucks, except for the whole brain emulation folks who might or might not be making progress
15:19  * nsh_ shrugs
15:20 < nsh_> i don't believe in artifice :)
15:20 < kanzure> nsh_: i've been geeking out about http://diyhpl.us/~bryan/papers2/neuro/ultrasound/
15:21 < nsh_> kanzure, what are you imagining to do with ultrasound?
15:21 < nsh_> general biochemical tinkering?
15:21 < kanzure> transcranial ultrasound for stimulation and/or brain melting
15:22 < nsh_> mmm
15:30 < nsh_> will look into this
15:30 < nsh_> bbl
15:34 -!- nsh_ [~nsh@cpc3-broo4-0-0-cust997.14-2.cable.virginmedia.com] has quit [Ping timeout: 276 seconds]
15:53 -!- PixelScum [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has joined ##hplusroadmap
15:56 -!- Drakkar [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has quit [Ping timeout: 252 seconds]
16:02 -!- eridu [~eridu@gateway/tor-sasl/eridu] has quit [Ping timeout: 250 seconds]
16:11 -!- klafka [~textual@cpe-69-205-70-55.rochester.res.rr.com] has joined ##hplusroadmap
16:16 -!- Avarice [~nsh@cpc1-cowc1-0-0-cust844.14-2.cable.virginmedia.com] has joined ##hplusroadmap
16:19 -!- Avarice [~nsh@cpc1-cowc1-0-0-cust844.14-2.cable.virginmedia.com] has quit [Client Quit]
16:20 -!- Avarice [~nsh@cpc1-cowc1-0-0-cust844.14-2.cable.virginmedia.com] has joined ##hplusroadmap
16:20 -!- Avarice is now known as nsh_
16:22 -!- nsh_ [~nsh@cpc1-cowc1-0-0-cust844.14-2.cable.virginmedia.com] has quit [Changing host]
16:22 -!- nsh_ [~nsh@wikipedia/nsh] has joined ##hplusroadmap
16:22 -!- nsh_ is now known as nsh
16:27 -!- alystair [alystair@24-246-14-18.cable.teksavvy.com] has joined ##hplusroadmap
16:44 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has quit [Read error: Connection reset by peer]
17:10 < jrayhawk> "< aristid> instead they should grant tax exemptions." huh, that's not a bad idea
17:12 < kanzure> seems marginally better than my "immediate nullification of all patents" plan
17:18 < QuantumG> you just want it both ways.  Don't like the restriction of patents, don't respect the necessity for trade secrets without patents.
17:19 < kanzure> every product should have its information/recipe available
17:22 < QuantumG> there ya go.
17:24 -!- eridu [~eridu@gateway/tor-sasl/eridu] has joined ##hplusroadmap
17:41 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has joined ##hplusroadmap
17:46 < kanzure> emokit w/ websockets http://makemyactionschainreactions.net/eeg/
18:39 -!- PixelScum [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has quit [Read error: Connection reset by peer]
18:39 -!- PixelScum [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has joined ##hplusroadmap
18:43 < fenn> intellectual property tax makes a lot of sense
18:43 < fenn> if we're going to go down the IP road
18:44 < fenn> or secrecy tax, whatever
19:00 < kanzure> haha if all your other property can be taxed.. why not IP :P
19:00 < kanzure> but unfortunately this will be misconstrued as something to tack on, not as an alternative to limited monopolies
19:00 < kanzure> (tax incentives vs. intellectual property tax)
19:04 < augur> kanzure: so no. there is no AIWG here. shame!
19:04 < kanzure> aiwg?
19:04 < kanzure> we have people that get excited about whole brain emulation,
19:04 < kanzure> i think that's close enough.
19:05 < augur> hrmph.
19:08 < kanzure> what would you want
19:11 -!- PixelScum [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has quit [Read error: Connection reset by peer]
19:11 -!- PixelScum [~PixelScum@ip98-177-175-88.ph.ph.cox.net] has joined ##hplusroadmap
19:34 < kanzure> is stross' "rule 34" really about rule 34?
19:39 -!- eridu [~eridu@gateway/tor-sasl/eridu] has quit [Ping timeout: 250 seconds]
19:41 < nsh> askance
19:42 < kanzure> is that a dance style
19:43 < nsh> it's a rare spice, depreciated in modern cuisine
20:12 -!- mrtrousers [5339b5dd@gateway/web/freenode/ip.83.57.181.221] has joined ##hplusroadmap
20:15 -!- jennicide [jen@173-29-174-114.client.mchsi.com] has quit [Read error: Connection reset by peer]
20:15 -!- jennicide [~jen@173-29-174-114.client.mchsi.com] has joined ##hplusroadmap
20:16 -!- AJollyLife [~quassel@unaffiliated/ajollylife] has quit [Ping timeout: 258 seconds]
20:20 < foucist> she glanced at his penis in askance
20:21 < kanzure> foucist: RAGE
20:21 < kanzure> http://heybryan.org/shots/2011-06-13-2249-arenasnow-typeracer.png
20:22 < kanzure> <-- raped :(
20:25 < foucist> just an example of proper usage
20:25 < foucist> of the word
20:25 < foucist> in a sentence
20:25 < foucist> heh
20:25 < kanzure> ?
20:25 < foucist> kanzure: someone beat you? about time
20:25 < foucist> (askance)
20:35 -!- jenzebubble [jen@173-29-174-114.client.mchsi.com] has joined ##hplusroadmap
20:38 -!- jennicide [~jen@173-29-174-114.client.mchsi.com] has quit [Ping timeout: 260 seconds]
20:40 < fenn> "The other element is found in the stenotype, that somewhat disconcerting device encountered usually at public meetings. A girl strokes its keys languidly and looks about the room and sometimes at the speaker with a disquieting gaze. Combine these two elements, let the Vocoder run the stenotype, and the result is a machine which types when talked to."
20:41 < fenn> 197 wpm, is that with a querty keyboard?
20:44 < kanzure> yes
20:44 < kanzure> the dude has arms of steel
20:44 < kanzure> http://www.youtube.com/watch?v=uf0hrleDE-E&NR=1
20:49 < fenn> that the input box clears every word makes comparison difficult
20:51  * kanzure is facebook friends with sean :(
20:51 < kanzure> if you can't beat him.. join him
20:52 < fenn> typaholics anonymous
20:53 < kanzure> "But my true obsession in high school was being statistician of the Mensa Scrabble-by-Mail SIG, where I was also a top ten player before I stopped playing upon entering college"
20:53 < kanzure> what
20:53 < fenn> do intelligent people use bigger words?
20:53 < fenn> calculemus!
20:54 < kanzure> 'mensa scrabble-by-mail SIG'
20:54 < kanzure> i can't stop laughing :(
20:54 < fenn> i was reading julian assange's blog earlier today
20:54 < fenn> he has this whole long diatribe about how intelligent people are socially outcast because they can't communicate with normals
20:55 < fenn> referring specifically to 150+ iq
20:55 < kanzure> bah my wpm is higher than that iq!
20:55 < kanzure> (i don't know what this means)
20:55 < fenn> i dont think i'd have the patience necessary for x-by-mail games
20:56 < kanzure> hey does 'units' handle iq
20:57 < fenn> no
20:57 < fenn> iq is based on percentile relative to the population at large
20:57 < fenn> the average is always exactly 100
20:58  * nsh mumbles something about relative-state
20:58 < fenn> units doesn't do things that change arbitrarily, except for some currency conversions (which i think i s a bad idea anyway)
21:00 < fenn> "almost all IQ tests adhere to the assignment of 15 IQ points to each standard deviation"
21:01 < fenn> it would be much better if they just said the mental age, me thinks
21:01 < kanzure> so our friend the 'smartest man in the world' is, what, 12 standard deviations out?
21:01 < kanzure> unlikely..
21:02 < fenn> the test breaks down at high end because there arent enough to draw a baseline from when designing the test
21:03 < fenn> if you get 100 out of 100 questions right, what's your score?
21:04 < fenn> hm, this is a long wikipedia entry
21:04 < ybit> aiwg?
21:04 < ybit> augur^
21:06 < augur> O_O
21:06 < ybit> don't give me that face
21:06 < augur> ybit: AI working group
21:06 < ybit> explain yourself :)
21:06 < ybit> okay
21:08 < fenn> "The accepted best measure of g is Raven's Progressive Matrices which is a test of visual reasoning."
21:08 < fenn> is it just me or does that statement not make any sense
21:08 < fenn> g is supposed to be "general" not "visual"
21:09 < fenn> i guess it's accepted because it's culturally neutral
21:09 < augur> ybit: i want to work on a project to design weak AGI around more formal techniques, but atypical ones
21:09 < augur> so i was curious if there was a wg here i could poke at
21:10 < ybit> #opencog
21:10 < fenn> augur: i want to pooh pooh your idea but you haven't told us enough
21:11 < augur> fenn: im not aiming for anything strong, just something fairly agent-ish that can engage relatively fluidly in natural language.
21:11 < fenn> but seriously i dont think any good will come of a system that does not use advanced statistics and massively parallel computation
21:11 < augur> the intention isnt for anything major, fenn.
21:11 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has quit [Ping timeout: 260 seconds]
21:12 < augur> i mean, honestly, i dont think that statistics will help, and im not convince that parallelism is truly necessary, but
21:12 < augur> thats beside the point
21:12 < fenn> ... One must think "divergently" of many possible solutions to a given, often ill-defined, problem; the relative scarcity or abundance of various resources, present and projected individual or group needs...
21:13 < augur> sure thats fine. it doesnt necessitate parallelism tho
21:13 < fenn> what is the intention then?
21:13 < augur> i mean, im all for parallelism. its computationally quite nice, if you can get it to work
21:14 < fenn> megahal engages fluidly in natural language... it just doesn't say anything sensical most of the time
21:14 < augur> but its not magically more powerful. its equivalent to serial processing, so i see the parallelism issue as a distraction.
21:14 < augur> brainlike "parallelism" on the other hand is a different issue
21:14 < augur> ive never heard of megahal. ill check it out.
21:14 < fenn> it's only equivalent to serial processing if the separate threads don't interact at all
21:14 < augur> oh yes, one of these things.
21:15 < fenn> markov model irc bot
21:15 < augur> fenn: no, serial machines can simulate any parallel machine, so they're completely equivalent.
21:15 < augur> parallelism is just useful for certain problem structures.
21:16 < augur> further, interaction of threads is a concurrency issue not a parallelism issue. tho the two are often conflated, they're quite distinct.
21:16 < fenn> and happens to be the only progress in computers at the moment
21:16 < augur> http://existentialtype.wordpress.com/2011/03/17/parallelism-is-not-concurrency/
21:16 < augur> the only progress how
21:17 < foucist> topological computation is the future
21:17 < foucist> fixes all the problems w/ parallelism
21:17 < fenn> there hasn't been significant improvement in clock speeds for years
21:17 < foucist> for certain classes of problems anyways
21:17 < augur> foucist: lazy/non-strict computation > all!
21:17 < augur> haskell will dominate!
21:18 < augur> fenn: oh thats true, yes. i dont deny that parallelism is a huge thing in computation right now.
21:18 -!- klafka [~textual@cpe-69-205-70-55.rochester.res.rr.com] has quit [Quit: Computer has gone to sleep.]
21:19 < augur> but thats a concern that should come /after/ the design of a correct program, not before. get the program to run correctly, then get it to run fast.
21:20 < augur> also, fenn, graphene processors promise to bump clockspeed to the 300-500GHz range, if recent news is accurate
21:20 < fenn> that article (parallelism is not concurrency) is not making sense to me
21:20 < augur> shame. i dont think i could explain it better than he does
21:20 < fenn> "parallelism is an abstraction, not an implementation" huh?
21:21 < fenn> i think he should start off with some definitions after "rebooting the reader"
21:22 < augur> also, fenn, if megahal is a markov model, i can tell you /precisely/ why it sucs
21:22 < fenn> because it's stateless
21:22 < augur> well besides that
21:22 < augur> there are many reasons why megahal sucks
21:22 < fenn> i think it's quite fun actually
21:23 < augur> the first, which statelessness relates to, is that natural language is not regular, probably not even context-free. markov models are strictly regular.
21:23 < fenn> tell me why it sucks more than any other simple AI algorithm
21:23 < augur> so its no surprise it produes garbled crap.
21:23 < fenn> you never answered what the intention was
21:24 < augur> secondly, its not really AI: its just chewing through text and running correlations to spit out new text, and would work similarly on, say, chess moves or musical notation or digestive patterns. it has no recourse to any sort of semantics to structure its behavior in any fashion.
21:24 < augur> i did tell you my intention!
21:25 < augur> something agent-ish!
21:25 < fenn> "its just chewing through text and running correlations to spit out new text"  is not intelligence?
21:25 < augur> well, you know, in as much as anything is intelligent :)
21:26 < fenn> i believe most humans do the same sort of thing, but have better filtering
21:26 < augur> i wouldnt consider just running statistical analyses of symbol transitions by itself to be enough for something to be considered AI
21:26 < augur> humans definitely Do do similar sorts of things
21:26 < augur> but its almost certainly not _all_ we do
21:26 < fenn> n-gram transitions perhaps
21:27 < fenn> again, this is getting lost in a text-based view of reality, which it certainly is not
21:27 < augur> i mean, at some level maybe its transitions of some sort. some crazy higher-order transition stuff who knows
21:27 < augur> thats also another problem -- megahal is just a text widget.
21:28 < fenn> and you expect your AI to make sense of non-textual data?
21:28 < fenn> suddenly it's not a simple problem
21:29 < augur> i dont expect it to make sense of anything significant. like i said, its intended to be an agent, so its non-textual data is just its affordances, its in-computer environment, etc.
21:31 < kanzure> wheee isn't this fun
21:31 < kanzure> also, i told you so
21:31 < fenn> kanzure: wah.
21:32 < kanzure> it's okay you have a chance to redeem yourself in robot hell
21:32 < fenn> augur: you should do something with music, real-time sound analysis and repetition. lots of room for creativity
21:32 < kanzure> what's wrong with whole brain emulation?
21:32 < fenn> it's too slow
21:33 < kanzure> in the scheme of things that's not a big problem
21:33 < augur> kanzure: nothing. its just not my goal right now. :)
21:33 < kanzure> i thought your goal was 'ai'
21:33 < kanzure> (and i thought my goal was goalism or something)
21:33 < kanzure> goal goal goal goal
21:33 < fenn> kanzure: would you like to play kurzweil for a day? when will we have sufficient computing power to simulate a human brain in real time for $1000?
21:33 < augur> also, i want to understand thought, not merely recreate it somehow
21:33 < augur> black boxes do not satisfy me
21:33 < kanzure> fenn: umm, uhh, doesn't theuncertainfuture.com do that
21:34 < fenn> yeah but your specific prediction
21:34 < kanzure> in fact, i'm pretty sure there's a few scenarios on there specifically for wbe
21:34 < augur> fenn: i have some minor interest in music stuff
21:34 < kanzure> well every time i put numbers into their it says a singularity is happening in 2 or 3 years
21:34 < augur> but not anything related to analysis, etc.
21:34 < augur> im more interested in formal theories
21:34 < fenn> formal theories are all bullshit
21:34 < kanzure> my mom was a formal theory
21:34 < kanzure> i can confirm this
21:34 < kanzure> former formal theory
21:35 < fenn> <- scientist, discordian
21:35 < fenn> augur: there's not enough to work with if you just feed your baby curated text
21:35 < kanzure> doesn't ENKI do that
21:36 < fenn> uh, too many enki's
21:36 < kanzure> ENKI-][:
21:36 < kanzure> that one.
21:36 < fenn> resolve name conflict please!
21:36 < kanzure> megahal namcub accela guy
21:36 < kanzure> john ohno
21:37 < fenn> until a bot shows up here and has a nice conversation here with me, i'll continue persisting in my beliefs about what will and will not work
21:37 < augur> fenn: uh, my intention _isnt_ to feed it curated text of any real sort. not for training.
21:37 < fenn> what then?
21:38 < augur> also, formal theories arent all bullshit. they work plenty well.
21:38 < fenn> the "if p then q" type don't work
21:38 < fenn> the "this looks like a p, maybe q" stand a chance
21:38 < augur> thats (classical) logic.
21:39 < augur> the majority of formal theories are  /not/ classical logic.
21:39 < fenn> do enlighten
21:39 < kanzure> pron.com user db? wtf http://lulzsecurity.com/releases/pronz.txt
21:39 < augur> well, when it comes to AI i cant say, actually
21:40 < kanzure> sorry but this still doesn't sound productive at all
21:40 < augur> but for logic itself, there are plenty of non-classical logics. linear logic and its descendants are pretty interesting. one is a logic of "causality" that has some interesting capacities.
21:40 < augur> frame-based reasoning works really well at text comprehension, at least from what ive seen.
21:41 -!- eridu [~eridu@gateway/tor-sasl/eridu] has joined ##hplusroadmap
21:41 < augur> and that was back in the 70s
21:41 < augur> hey eridu.
21:41 < eridu> hello augur
21:42 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has joined ##hplusroadmap
21:42 < augur> hows life, dude
21:42 < eridu> busy as fuck
21:43 < fenn> the wp article on frame reasoning mentions "spreading activation" which sounds a lot like neural networks to me
21:43 < fenn> are you saying neural networks are a formal theory?
21:44 < augur> eridu: anything good?
21:44 < eridu> not really
21:44 < eridu> lot of half-finished projects and lost campaigns
21:45 < eridu> I guess I got published in between now and when we last talked
21:45 < augur> fenn: spreading activation in the frame context just means that there is a semantic network that structures the frames, whereby the frame problem is solved by biased search to the activated frames
21:46 < augur> fenn: eg frame 1 "activates" frame 2, which means when the program looks for new frames to apply to the situation, itll check frame 2 first
21:46 < augur> thats a toy explanation of spreading activation but you get the point
21:46 < augur> the frames are just formal objects akin to predicates with more information
21:46 < augur> eridu: oh? published for?
21:47 < augur> eridu: also, lost campaigns? :\
21:47 < augur> did you run for emperor?
21:47 < fenn> what creates the bias for frame 2 instead of some other frame?
21:47 < kanzure> frame.. context.. words, bitches
21:47 < fenn> also my google fu is failing at "logic of causality"
21:48 < eridu> augur: yeah, the administration fucked us over on daycon
21:48 < eridu> also sds is dead
21:48 < fenn> http://en.wikipedia.org/wiki/Causality#Logic  looks like "if p then q" logic to me
21:49 < augur> fenn: the background semantic network that the designers built in. things like restaurant frames are associated with eating frames, paying frames, etc. but not with, say, sky diving frames
21:49 < augur> fenn: you'd think so, but thats because natural language is tricking you ;)
21:49 < augur> fenn: classical implication (if-then) is strictly a correlational property, not a causational one.
21:49 -!- eridu [~eridu@gateway/tor-sasl/eridu] has quit [Quit: Leaving]
21:50 < augur> actually natural language's conditional is also correlational, but there's a strong pragmatic implicature in certain contexts
21:50 < augur> there are many ways to override this tho.
21:50 < augur> i can find you the causal logic thing. they address AI-related issues in it, actually.
21:50 < augur> classical ones tho, nothing magical.
21:51 < augur> Prolegomena of a Logic of Causality and Dynamism - Bellot et al.
21:51 < kanzure> i can never really size up people when talking about ai
21:51 < kanzure> i have no idea if you're a machine learning newbie or not
21:51 < kanzure> but i do remember you wrote some linguistics papers?
21:51 < augur> im a linguist, yes
21:51 < augur> and im familiar with many ideas form ML, i just dont know jack about the details
21:52 < augur> fenn: http://wellnowwhat.net/transfers/Prolegomena%20of%20a%20Logic%20of%20Causality%20and%20Dynamism%20-%20Bellot%20et%20al..pdf
21:53 < fenn> ML is just a fancy word for "statistics with computers"
21:54 < kanzure> search algorithms are statistics now?
21:54 < fenn> most people think of statistics as what they learned in high school, so it really does need to be sexed up
21:54 < fenn> yeap pretty much
21:55 < kanzure> the mlai world just keeps redefining shit in a million ways
21:55 < kanzure> this is really useless guys..
21:55 < augur> kanzure: ML is more than just search algos
21:55 < augur> well, depends on what you mean by search i guess
21:55 < kanzure> yes i know
21:56 < augur> ultimately, fenn, ofcourse statistical inferences need to be used
21:57 < augur> but you cant do statistics without a good underlying model that is non-statistical
21:57 < fenn> yes we need to implement it after all
21:58 < fenn> have you ever read "reciprocality"?
21:58 < augur> nope
21:58 -!- mayko [~mayko@71-22-217-151.gar.clearwire-wmx.net] has quit [Ping timeout: 240 seconds]
21:58 < augur> statistics is an addition to an underlying formal theory tho, i must stress this
21:58 < fenn> http://www.buildfreedom.com/content/reciprocality/
21:59 < fenn> formal theories are what you do with the last step, once you know it's probably a cat and not a chair
21:59 < augur> not at all true
21:59 < fenn> the problem seems to be determining it's a cat
21:59 < augur> you simply cant do any statistics without an underlying non-statistical theory
21:59 < augur> simply because without that, theres nothing to run statistics over!
21:59 < fenn> yeah sure, like "this is a matrix with pixel values from a camera"
22:00 < fenn> that's not a theory, that's a statement of fact
22:00 < augur> yeah but what statistics will you calculate off of that
22:00 < fenn> i'd run lots of things in parallel
22:00 < augur> yeah but what statistics
22:00 < fenn> correlation between feature clusters
22:01 < augur> feature clusters are not in the pixel values
22:01 < fenn> right
22:01 < augur> congratulations, youve just invented a theory
22:01 < augur> and a formal one at that
22:01 < fenn> i am underwhelmed
22:01 < augur> because the feature clusters are below the stats!
22:01 < augur> you should be underwhelmed!
22:02 < augur> vision theory is a different issue tho
22:02 < augur> ofcourse AI should have a statistical vision theory underlying it
22:02 < fenn> (i mean feature as in the SIFT points, not in the statistics sense)
22:02 < augur> but vision theory itself, as a whole, has the stats on the top.
22:03 < augur> if you stratify to separate concerns, you get the formal theory at the bottom of each module, with stats on top of that
22:03 < augur> stats over a theory of the visual system, stats over a theory of sentence structure, stats over a theory of world knowledge, stats over a theory of human behavior, etc.
22:04 < fenn> i don't like hard coding theories
22:04 < augur> you have to have _something_ to calculate stats over. even if its just your theory of visual features.
22:05 < fenn> currently with limited computer power can only perform brittle crystalline calculations on pixel values, but it doesn't *have* to be that way
22:05 < augur> or you theory of theories of visual features, if you want to run higher-order statistics
22:05 < augur> i agree completely.
22:05 < augur> but you still have to have a theory of what matters and what doesnt, what is looked for and what isnt
22:05 < fenn> if you say so
22:06 < augur> the intuitive stuff like edges, polygons, shading, etc. are all sensible, ofcourse
22:06 < augur> but those arent derived from the stats
22:06 < augur> they precede it
22:06 < fenn> "of course" is what killed AI research
22:06 < augur> ;)
22:06 < epitron> nature doesn't have many polygons :(
22:06 < fenn> "of course space is euclidean"
22:07 < augur> epitron: sure it does, they're just really tiny ;)
22:07 < epitron> and in terms of how the brain works, polygons are the highest level
22:07 < augur> fenn: well, afaact it IS euclidean!
22:07 < epitron> they're very abstract and generalized
22:07 < epitron> like language
22:07 < epitron> your modules are upside-down
22:07 < epitron> you're doing top-down plus bottom up statistics
22:07 < augur> epitron: im not saying polys really are at the bottom
22:08 < augur> i was using that as an example
22:08 < epitron> sure
22:08 < epitron> but it's still upside-down :)
22:08 < augur> perhaps!
22:08 < augur> my point is that you cant run statistics over nothing
22:09 < fenn> the egyptians drew their maps with the source of the nile at the top
22:09 < augur> you always have to make decisions about what the underlying theory is.
22:09 < augur> and that underlying theory is not itself a statistical one
22:09 < epitron> you mean, assumptions to limit your search space?
22:09 < epitron> those are good
22:09 < fenn> he just means how you determine the set of data to do statistics upon
22:09 < epitron> well, if they're good assumptions
22:09 < epitron> true
22:10 < epitron> good data is lso good :)
22:10 < fenn> it's not even a search space at that point
22:10 < augur> fenn: often enough the data you run stats on is structured in some way or other
22:10 < fenn> let's just say it's binary data
22:10 < fenn> what regularities do you see?
22:10 < epitron> i wonder how many people in the world are talking about re-engineering the brain right now
22:10 < epitron> probably a lot
22:10 < augur> fenn: that depends on what sort of grammar you're using!
22:10 < epitron> at least a few hundred
22:10 < fenn> can you infer ascii text based on the absence of a 1 in the 8th bit?
22:10 < augur> if you're using a 1-gram model, you get jack shit
22:11 < augur> if you're using a 10-gram model you get more
22:11 < augur> if you allow context free behavior, you get even more
22:11  * epitron rolls a 1-gram model
22:12 < fenn> a few hundred is not a lot compared to most fields
22:12 < augur> fenn: you have to make a choice even in binary processing about what sort of structure the data is assumed to have
22:13 < augur> if you choose a CF model, you will get better results than if you have a regular model. that is a fact.
22:13 < augur> at least if the data is CF in structure
22:13 < fenn> CF?
22:13 < augur> context-free
22:14 < fenn> if you start with the answer, you will get a better answer
22:14 < augur> a classical example of a CF language, in pure binary, would be: all bit strings with a bunch of 0's at the beginning, and the same number of 1's folloiwng that
22:15 < augur> a markov model simply can never learn this. it might _MIGHT_ be able to get arbitrarily close. but that iwll also depend on how many states you allow it to have, etc. etc.
22:15 < augur> fenn: im not starting with the answer tho thats the point
22:15 < kanzure> 'america invents act' http://www.hoover.org/publications/defining-ideas/article/82096
22:15 < augur> youre always making assumptions about the nature of the system. assumptions which arent based on statistics.
22:16 < augur> sure you can generalize your system so that it can dynamically shift from one kind of assumption to the next, or whatever, but you're not escaping the underlying non-statistical nature of the system
22:16 < fenn> you can start with a large number of statistical models and choose the one that predicts the data best
22:17 < augur> all you're doing is making the non-statistical component adaptive
22:17 < augur> fenn: sure, but those statistical models are still statistics over SOMETHING
22:17 < augur> and that something is not statistical in nature
22:17 < fenn> noisy sensor data
22:17 < augur> not just
22:17 < fenn> don't get all "digital sampling is a formal theory" on me
22:18 < augur> your stats are stats of SOMETHING
22:18 < augur> state transitions, production rules, etc.
22:18 < augur> they're stats running over something non-statistical
22:18 < fenn> is a fourier transform a formal theory?
22:19 < augur> that depends on your opinion of math. it is certainly a purely mathematical concept.
22:19 < augur> but its also tangential to the point
22:19 < augur> you run a fourier transform, but /then what/?
22:19 < augur> now you just have data of a different shape.
22:19 < augur> big deal.
22:20 < augur> what are you attributing probabilities to, fenn
22:20 < augur> probabilites are probabilities of _something_
22:20 < fenn> it is an assumption to make in order to get better predictability out of your statistics
22:20 < augur> sure, fourier transforms are nicer to deal with. i know that
22:20 < augur> but that doesnt change anything
22:20 < augur> what are you calculating the probabilities of. thats the point.
22:20 < augur> what has the probabilities.
22:21 < fenn> arbitrary things
22:21 < fenn> i dont know the word
22:21 < fenn> a variable
22:21 < augur> i feel like you're not very familiar with machine learning...
22:21 < fenn> monica calls them 'hypotheses'
22:22 < augur> well i dont know what monica is
22:22 < fenn> a person who does ai research
22:22 < augur> yyyeah...
22:23 < fenn> different algorithms have different sorts of outputs, it's not always straightforward to integrate many different algorithms
22:23 < kanzure> are you two going to blather about bayesian bullshit yet
22:23 < kanzure> *poke*
22:23 < augur> look, its clear you're not familiar enough with ML techniques to even defend your own position adequately
22:23  * kanzure pops some popcorn
22:23 < QuantumG> sooo.. linguistics eh.. are dictionaries descriptive or prescriptive?
22:23 < augur> kanzure: well, bayesian inference is ofcourse superior, so.
22:24 < augur> QuantumG: depends on the dictionary.
22:24 < QuantumG> can a prescriptive dictionary be the result of linguistics?
22:24 < augur> no, but i would argue that descriptive ones cant either.
22:24 < fenn> if it's a conlang
22:25 < QuantumG> do tell
22:25 < augur> linguistics is the scientific study of language
22:25 < augur> that immediately rules out prescriptivism in any form
22:25 < QuantumG> agreed
22:25 < augur> science describes, models, etc. it does not proscribe
22:26 < QuantumG> go on
22:26 < augur> further, dictionaries are documentations of usage. documentation is important, but its not study. there is no explanation to be done with dictionaries.
22:27 < QuantumG> ahh.. I see your argument.  Well, intermediate products are also products.
22:27 < augur> they are, but they're not necessarily the same sort of thing as the end product.
22:28 < augur> You need gasoline for cars, but oil refinement is not automotive engineering.
22:28 < augur> i have great respect for lexicographers, but its not linguistics.
22:28 < augur> there is no science being done there.
22:29 < fenn> i dont see how you can do science by making formal models of a system that routinely breaks the rules
22:30 < augur> you'd be surprised how little language actually violates the rules.
22:30 < augur> you just have to know what the rules actually are.
22:30 < QuantumG> discovering that there are any rules at all is the greatest achievement of linguistics
22:30 < fenn> meh
22:31 < augur> meh!
22:31 < fenn> grammar goes back thousands of years
22:31 < augur> yes it does, but not as a science
22:31 < augur> grammar as a prescriptive act certainly does tho.
22:31 < QuantumG> physics goes back thousands of years.. science is the method of refinement
22:31 < fenn> which came first, the language or the prescriptivist grammar?
22:32 < augur> the language, surely.
22:32 < augur> you cant be snobby about french if there aint no french!
22:32 < fenn> so where did the grammar come from?
22:32 < augur> grammar is in peoples heads.
22:32 < QuantumG> physical reality of course.
22:33 < augur> writing down a grammar for a language is just an attempt to capture the knowledge people have in their heads
22:33 < QuantumG> linguistics is as much about the study of physical reality as physics.
22:33 < fenn> some people would say genetics or evolution, but i'm not sure
22:33 < augur> fenn: thats where the FACULTY of language comes from
22:33 < fenn> oh the FACULTY
22:33 < augur> but the grammars themselves are not genetic/evolved
22:33 < fenn> that makes it so clear!
22:33 < augur> fenn: think of it like this
22:34 < augur> being able to move around on two legs coordinatedly is genetic
22:34 < augur> waltzes are not.
22:34 < fenn> feral children raised by wolves do not walk on two legs
22:34 < augur> this is (perhaps?) true
22:34 < augur> the brain is not static
22:34 < fenn> also there are dogs with only two hind legs that balance
22:34 < augur> unused portions of the brain are reallocated.
22:35 < augur> we know this.
22:35 < fenn> if all humans walked on four legs, so would their children
22:35 < augur> probably!
22:35 < fenn> what makes waltzes less genetic than walking?
22:35 < augur> but take one of their children and raise the kid among bipeds, and the child would not have trouble learning it competently
22:36 < augur> and without any training, too
22:36 < augur> but a dog will not.
22:36 < fenn> let us imagine a machine that can perform any possible sequence of foot maneuvers
22:36 < augur> admittedly they probably have the cognitive capacity, its mostly in their physical structure thats an issue
22:36 < augur> they also probably lack the finer balance stuff that bipeds have
22:36 < fenn> i bet you a bajillion dollars most humans will not be able to learn a significant portion of its dances
22:36 < augur> certainly
22:37 < augur> its an analogy, fenn
22:37 < QuantumG> not that it's the topic, but I don't think we'd consider them "dances" .. just as we don't consider many sequences of sounds to be language.
22:37 < augur> point is, languages are not genetically encoded. the ability to use them is.
22:37 < fenn> i disagree
22:37 < augur> well you're wrong.
22:38 < fenn> it has been a displeasure talking with you
22:38 < augur> the feeling is mutual, o lay friend.
22:38  * kanzure eats more popcorn
22:38 < kanzure> when will you mortals learn that i am the superior option
22:39 < fenn> what's the inverse of "told you so"?
22:39 < kanzure> "you sure told me"?
22:39 < augur> fenn: "you're right"
22:39 < augur> or "you were right"
22:53 < fenn> btw, space is not euclidean, in case that was in doubt.
22:54 < augur> all measurements show it to be flat, fenn.
22:54 < augur> there's no detectable curvature
22:54 < augur> you know, modulo gravity
22:55 < QuantumG> all evidence indications that space is euclidean
22:55 < augur> the large-scale geometry of space is flat. locally, gravity changes things ofcourse.
23:00  * augur steals kanzure's popcorn
23:02 -!- mode/##hplusroadmap [+o kanzure] by ChanServ
23:02 -!- augur was kicked from ##hplusroadmap by kanzure [augur]
23:02 <@kanzure> don't fucking steal my motherfucking popcorn
23:02 < QuantumG> heh
23:02 -!- mode/##hplusroadmap [-o kanzure] by ChanServ
23:03 -!- augur [~augur@208.58.6.161] has joined ##hplusroadmap
23:03 < augur> :|
23:03  * augur knocks kanzure's popcorn out of his hands
23:03 < augur> >|
23:08 < fenn> really, there arent any articles written about "culture invariance of grammar"?
23:09 < augur> sorry?
23:10 < augur> what are you looking for, i can direct you
23:10 < fenn> what is the verbal equivalent of phosphenes called?
23:10 < augur> what
23:10 < fenn> characteristics of grammar that are preserved across all cultures
23:10 < QuantumG> culture invariance of grammar = universal grammar
23:10 < QuantumG> there's an absurd number of articles about it.
23:10 < augur> i dont think that phosphenes are what you mean
23:11 < augur> but yes, grammatical universals are the property in question
23:11 < augur> universal grammar is the theory of grammatical universals
23:12 < augur> maybe you're thinking of form constants, fenn?
23:12 < QuantumG> study of how children "flip the bits" on the universal grammar to learn the particular native dialect is about a year's worth of reading.
23:12 < augur> honestly, im skeptical of that sort of parametricity, QuantumG.
23:12 < augur> ive never bought it, to be honest.
23:12 < fenn> no, not form constants
23:13 < fenn> phosphenes are patterns and colors you see when you push on your eyeballs
23:13 < fenn> all humans see the same patterns
23:13 < augur> perhaps
23:14 < augur> if so, only because of the structure of the eye
23:14 < augur> form constants are a better analogy to universals.
23:14 -!- jenzebubble [jen@173-29-174-114.client.mchsi.com] has quit [Ping timeout: 258 seconds]
23:14 < QuantumG> which is the same argument for universal grammar.  That we all have the same machinery for "looking" at language.
23:15 < augur> actually if there were an eye-ball analogy it would relate to ears :p
23:16 < QuantumG> well, I was more thinking about edge detection and other visual cortex processing that we all do.
23:16 < fenn> the drawings online of form constants dont look anything like phosphene shapes to me
23:16 < QuantumG> but yes
23:16 < augur> fenn: no, they dont
23:16 < augur> they're not supposed to
23:16 < augur> form constants are shapes that are common to many sorts of atypical visual stimulation
23:17 < fenn> often phosphenes look like characters written in some language i can't read, or grid or moire patterns
23:17 < augur> eg hallucinogens
23:18 < augur> you have interesting phosphenes
23:18 < augur> ive never had that
23:19 < fenn> well, so much for that theory :P
23:22 < augur> what
23:22 < fenn> that all humans see the same patterns
23:23 < fenn> probably just me misremembering what i read
23:23 < augur> you read about form constants
23:23 < mrtrousers> Anyone interested in coming to Barcelona and get a Hackerspace started?
23:23 < mrtrousers> There will be biohacking!
23:23 < augur> mrtrousers: i wish i could!
23:29 < mrtrousers> ;)
23:49 -!- klord [~klord@99-67-239-104.lightspeed.austtx.sbcglobal.net] has quit [Ping timeout: 252 seconds]
23:57 -!- klord [~klord@99-67-239-104.lightspeed.austtx.sbcglobal.net] has joined ##hplusroadmap
--- Log closed Tue Jun 14 00:00:07 2011