| --- Day changed Sun Nov 30 2014 | ||
| archels | < kanzure> there should be an embodied turing test. | 00:38 |
|---|---|---|
| -!- ebowden [~ebowden@147.69.179.205] has quit [Read error: Connection reset by peer] | 00:38 | |
| -!- ebowden_ [~ebowden@147.69.179.205] has joined ##hplusroadmap | 00:38 | |
| archels | the Turing test is embodied insofar as any and all higher-level cognition is posited to be based on embodiment | 00:38 |
| archels | (although not by Turing) | 00:39 |
| archels | what were you thinking of--spatial nagivation/manipulation tasks? | 00:39 |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 00:43 | |
| -!- Burn_ [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has quit [Read error: Connection reset by peer] | 01:20 | |
| -!- Burn_ [~Burn@pool-71-191-174-26.washdc.fios.verizon.net] has joined ##hplusroadmap | 01:20 | |
| -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 01:30 | |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 256 seconds] | 01:49 | |
| delinquentme | http://www.theguardian.com/society/2014/nov/27/worlds-most-expensive-medicine-glybera-sale-1million-price | 01:53 |
| delinquentme | sez china has gene therapeutics ..... | 01:53 |
| -!- Guest17879 [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has quit [Remote host closed the connection] | 01:53 | |
| -!- maaku [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has joined ##hplusroadmap | 01:53 | |
| -!- maaku is now known as Guest10156 | 01:54 | |
| delinquentme | Gendicine | 01:54 |
| -!- HEx1 [~HEx@hexwab.plus.com] has quit [Ping timeout: 258 seconds] | 02:23 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 02:23 | |
| -!- Guest10156 [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has quit [Remote host closed the connection] | 02:26 | |
| -!- maaku_ [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has joined ##hplusroadmap | 02:26 | |
| -!- maaku_ [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has quit [Quit: No Ping reply in 180 seconds.] | 02:34 | |
| -!- maaku_ [~quassel@173-228-107-141.dsl.static.fusionbroadband.com] has joined ##hplusroadmap | 02:35 | |
| archels | .title http://arxiv.org/abs/1411.2731 | 03:03 |
| yoleaux | [1411.2731] Imaging at depth in tissue with a single-pixel camera | 03:03 |
| archels | 3 mm object, 2.4 mm depth | 03:04 |
| archels | hmm, 500 illumination patterns... they don't say how fast the image was acquired | 03:06 |
| archels | with DLP I imagine you could obtain it quite fast | 03:14 |
| jrayhawk | http://i.imgur.com/G8OIOPy.jpg how do you even do this in a pre-silkroad environment? | 03:18 |
| chris_99 | haha | 03:20 |
| -!- HEx1 [~HEx@hexwab.plus.com] has joined ##hplusroadmap | 03:35 | |
| jrayhawk | kanzure: do you have a way to make irssi do hierarchical /window management | 03:36 |
| jrayhawk | i am a little annoyed about losing hilights after window 90 | 03:37 |
| jrayhawk | and sometimes i want to filter activity flagging by network | 03:37 |
| -!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 264 seconds] | 03:41 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Remote host closed the connection] | 03:50 | |
| -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 04:14 | |
| -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 04:31 | |
| -!- Beatzebub_ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 04:37 | |
| -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Ping timeout: 256 seconds] | 04:41 | |
| -!- delinquentme [~dingo@74.61.157.78] has quit [Ping timeout: 255 seconds] | 04:45 | |
| -!- Beatzebub_ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Ping timeout: 240 seconds] | 04:46 | |
| -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 05:10 | |
| -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 244 seconds] | 05:12 | |
| -!- Viper168_ is now known as Viper168 | 05:13 | |
| -!- JayDugger [~jwdugger@pool-173-57-55-138.dllstx.fios.verizon.net] has quit [Quit: Leaving.] | 05:22 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 05:40 | |
| archels | http://massively.joystiq.com/2014/06/23/linden-lab-confirms-new-virtual-world-incompatible-with-second-l/ | 05:44 |
| -!- ebowden_ [~ebowden@147.69.179.205] has quit [Remote host closed the connection] | 06:03 | |
| kanzure | archels: spatial tasks, sure. but more like, "there is a human interacting with a robot and the human is trying to communicate and get the robot to do things" | 06:24 |
| kanzure | jrayhawk: no i don't | 06:24 |
| -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:6488:9f8a:b77:a230] has joined ##hplusroadmap | 06:35 | |
| -!- ThomasEgi [~thomas@2a02:810b:33f:dc18:6488:9f8a:b77:a230] has quit [Changing host] | 06:35 | |
| -!- ThomasEgi [~thomas@panda3d/ThomasEgi] has joined ##hplusroadmap | 06:35 | |
| -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-dyioaztnbawcewov] has joined ##hplusroadmap | 06:45 | |
| -!- Zinglon [~Zinglon@D549A77D.cm-10-1a.dynamic.ziggo.nl] has joined ##hplusroadmap | 06:50 | |
| -!- genehacker [~chatzilla@pool-173-57-37-204.dllstx.fios.verizon.net] has joined ##hplusroadmap | 06:56 | |
| -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has quit [Remote host closed the connection] | 07:05 | |
| -!- Merovoth [~Merovoth@gateway/tor-sasl/merovoth] has joined ##hplusroadmap | 07:20 | |
| kanzure | "Listen, we MAY have disassembled the LHC to build a death ray." "Just one, though." "Nothing you should worry about." "The death isn't even very serious." | 08:16 |
| heath | jrayhawk: zoom out of your terminal if you are wondering if you have more hilights | 08:32 |
| heath | otherwise, use weechat | 08:32 |
| kanzure | ha zooming | 08:33 |
| heath | the only other way is there's a shortcut in irssi for hopping to the next hilighted channel | 08:33 |
| kanzure | one alternative to implementing vertical lists would be to reserve the right-most side of the window bar to be for highlights | 08:34 |
| kanzure | and as more highlights happen they would expand leftwards until they take up the entire window bar (at which point you should probably deal with those highlights) | 08:34 |
| kanzure | any highlighted window that is already in the visible range should probably stay in the visible range and not be moved to the right | 08:35 |
| kanzure | don't i sound clever? | 08:35 |
| kanzure | nsh: so does one-tab.com store an archive of each page in cache, or is it just lists of links? | 08:36 |
| -!- sheena [~home@S0106c8be196316d1.ok.shawcable.net] has quit [Ping timeout: 256 seconds] | 08:44 | |
| -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-dyioaztnbawcewov] has quit [Quit: Connection closed for inactivity] | 08:47 | |
| -!- LennyKitty [~lk@unaffiliated/lennykitty] has joined ##hplusroadmap | 08:56 | |
| -!- LennyKitty [~lk@unaffiliated/lennykitty] has left ##hplusroadmap [] | 09:01 | |
| -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:8c51:9d75:8b6e:768] has joined ##hplusroadmap | 09:08 | |
| kanzure | http://www.vetmed.wsu.edu/depts-prd/programs/images/equineCT.jpg | 09:20 |
| chris_99 | wow | 09:21 |
| kanzure | http://www.artecimaging.com/horse1.jpg | 09:23 |
| -!- genehacker [~chatzilla@pool-173-57-37-204.dllstx.fios.verizon.net] has quit [Ping timeout: 245 seconds] | 09:31 | |
| -!- genehacker [~chatzilla@pool-173-57-37-204.dllstx.fios.verizon.net] has joined ##hplusroadmap | 10:08 | |
| kanzure | "Certainly after ten years of study we could say with confidence that carriers of the DAD2 receptor A1 allele have compromised D2 receptors. Therefore lack of D2 receptors causes individuals to have a high risk for multiple addictive, impulsive and compulsive behavioral propensities" | 10:12 |
| kanzure | "Blum united this hypodopaminergic trait under the rubric of a reward deficiency syndrome" | 10:12 |
| kanzure | "world inadequacy syndrome" | 10:12 |
| -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-udgtfhtywffecsup] has joined ##hplusroadmap | 10:18 | |
| -!- _Sol_ [~SolGr@c-69-141-24-242.hsd1.nj.comcast.net] has joined ##hplusroadmap | 10:55 | |
| -!- drewbot [~cinch@ec2-23-20-191-43.compute-1.amazonaws.com] has quit [Remote host closed the connection] | 11:27 | |
| -!- drewbot [~cinch@ec2-50-16-19-212.compute-1.amazonaws.com] has joined ##hplusroadmap | 11:28 | |
| kanzure | "Interesting article in Nature. Besides commenting on the incorrect equivalence of OA and online, I also commented on the increasing lack of accuracy and value due to decreasing attention to prior searching and the rise of “Google only” searching. ... 'As numbers of published articles rise, the scholarly review system must adapt to avoid unmanageable burdens and slipping standards, says Martijn Arns.'" | 11:35 |
| kanzure | "The publishing explosion used to be easy to illustrate by walking young scholars past cumulative indices of Chemical Abstracts and Science Citation Index. There have always been more demands on scholars' time than there are hours in the day." | 11:35 |
| kanzure | "Unlike the commentator, I've spoken to numerous faculty at my institution about this. More than a few have opted to stop reviewing for commercial publishers (Elsevier, Springer, Wiley, even NPG), and only accept requests to review for scholarly societies. A substantial number have made similar commitments with respect to their submissions, as well." | 11:35 |
| kanzure | maybe reviews should be split up more topically | 11:41 |
| kanzure | so when a reviewer hits a paper that uses a statistical test to draw a conclusion, he makes some review decisions assuming both that the test is accurate and then that the test is totally broken. | 11:41 |
| kanzure | then that segment gets sent off to someone who does statistics reviews more regularly. | 11:42 |
| kanzure | (or possibly to some generic software tool that reviews the use and applicability of various bad ideas, once common themes and problems are detected and known) | 11:42 |
| kanzure | although this wont stop the "everyone is spending all of their time reviewing and writing papers" problem | 11:43 |
| kanzure | another possible solution is to not have papers reviewed so readily, and instead wait until peers cite the papers in giant bundles and the bundles get reviewed all at once | 11:43 |
| kanzure | even if you decouple academic career metrics from publications i think you will still find this review overload problem | 11:44 |
| kanzure | oh i suppose the review problem is less bad in academia because if the reviewers say it's shit then they can kill it there | 11:46 |
| chris_99 | Anyone seen this AFM design before - http://www.kurzweilai.net/how-to-build-a-low-cost-afm-nanoscope-out-of-lego-arduino-board i can't see any instructions out there though | 12:03 |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Ping timeout: 258 seconds] | 12:08 | |
| -!- Zinglon [~Zinglon@D549A77D.cm-10-1a.dynamic.ziggo.nl] has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Chicks dig it] | 12:10 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 12:13 | |
| kanzure | "Classic experiments have shown that the desert ant keeps track of its position in this way as it wanders for many meters searching for food. Though it travels in a randomly twisted path, it heads straight home when it finds food. However, if the ant is picked up and released some meters to the east, for example, it heads for a location displaced by the same amount to the east of its home nest." | 12:15 |
| -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 12:16 | |
| -!- Russell [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 12:19 | |
| -!- Russell is now known as Guest1309 | 12:20 | |
| kanzure | "The sentience quotient concept was introduced by Robert A. Freitas Jr. in the late 1970s.[7] It defines sentience as the relationship between the information processing rate of each individual processing unit (neuron), the weight/size of a single unit, and the total number of processing units (expressed as mass). It was proposed as a measure for the sentience of all living beings and computers from a single neuron up to a hypothetical ... | 12:23 |
| kanzure | ... being at the theoretical computational limit of the entire universe. On a logarithmic scale it runs from −70 up to +50." | 12:23 |
| kanzure | http://en.wikipedia.org/wiki/Sentience_quotient | 12:23 |
| kanzure | i keep forgetting how old freitas must be :p | 12:23 |
| kanzure | "According to Freitas, an alien civilization having their consciousness running on non-biological hardware (such as quantum-mechanical circuits) could have an SQ of 23+, 10 orders of magnitude more than the human SQ. Freitas states that such a gap in SQ "may affect our ability, and the desirability, of communicating with extraterrestrial beings...It may be that there is a minimum SQ "communication gap," an intellectual distance beyond ... | 12:24 |
| kanzure | ... which no two entities can meaningfully converse." [1] For example, an alien civilization may form a Matrioshka brain or a black hole and communicate using neutrinos or gamma-ray bursts at bandwidths that exceed our receiving capabilities." | 12:24 |
| -!- Vutral [~ss@mirbsd/special/Vutral] has quit [Ping timeout: 245 seconds] | 12:26 | |
| -!- genehacker [~chatzilla@pool-173-57-37-204.dllstx.fios.verizon.net] has quit [Remote host closed the connection] | 12:28 | |
| jrayhawk | looks like the community solution to that irssi problem is to send all hilights to a dedicated window and just leave it on the screen via window splitting | 12:31 |
| kanzure | that sounds silly | 12:32 |
| kanzure | i like my "move highlights to the right side of the window bar" idea better | 12:32 |
| -!- Vutral [ss@mirbsd/special/Vutral] has joined ##hplusroadmap | 12:34 | |
| kanzure | freitas was born 1952? | 12:52 |
| -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has joined ##hplusroadmap | 13:02 | |
| kanzure | "One theory of attention-deficit hyperactivity disorder states that ADHD can lead to deficits in working memory.[59] Studies suggest that working memory can be improved by training in ADHD patients through computerized programs.[60] This random controlled study has found that a period of working memory training increases a range of cognitive abilities and increases IQ test scores. Consequently, this study supports previous findings ... | 13:08 |
| kanzure | ... suggesting that working memory underlies general intelligence." | 13:08 |
| kanzure | paperbot: http://www.tandfonline.com/doi/abs/10.1076/jcen.24.6.781.8395#.VHuHSlRGjUY | 13:08 |
| paperbot | http://libgen.org/scimag/get.php?doi=10.1076%2Fjcen.24.6.781.8395 | 13:08 |
| kanzure | aww a negative result http://www.gwern.net/docs/dnb/2012-chooi.pdf | 13:12 |
| -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 264 seconds] | 13:13 | |
| -!- snuffeluffegus [~snuff@ps357888.dreamhost.com] has joined ##hplusroadmap | 13:18 | |
| -!- Viper168 [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 13:21 | |
| kanzure | heh! "Previous research using other paradigms has shown that individuals high in WMC are not as likely to experience false memories as low-..." | 13:23 |
| kanzure | "This study investigated how to teach perceptual tasks, that is, classifying fish locomotion, through eye movement modeling examples (EMME). EMME consisted of a replay of eye movements of a didactically behaving domain expert (model), which had been recorded while he executed the task, superimposed onto the video stimulus" | 13:27 |
| kanzure | oh that is a good idea | 13:27 |
| kanzure | better than laser pointers | 13:27 |
| kanzure | "Our findings suggest that frontostriatal mechanisms track the utility of information in WM, and that these dynamics may predict delays in the removal of information from WM." | 13:33 |
| kanzure | "Relatedly, direct injections of a D1 agonist into the prelimbic cortex have been shown to enhance the performance of animals on working memory tasks" http://www.rci.rutgers.edu/~totto/Wass%20et%20al%202013.pdf | 13:52 |
| kanzure | from the same author file:///home/kanzure/Downloads/Matzel,%20CD,%202013%20PROOF%20(1).pdf | 13:55 |
| kanzure | man i hate reseachgate | 13:55 |
| kanzure | http://www.researchgate.net/publication/256114562_The_architecture_of_intelligence_Converging_Evidence_from_studies_of_humans_and_animals/file/e0b49521ca858c0dfc.pdf | 13:55 |
| heath | http://www.technologyreview.com/featuredstory/532691/google-glass-is-dead-long-live-smart-glasses/ | 14:09 |
| heath | http://www.innovega-inc.com/ | 14:09 |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 14:09 | |
| heath | "A company called Innovega is doing this by developing contact lenses with a tiny bump that serves as a microscope for content that can be streamed from the inside of a pair of glasses. The lenses do nothing when you’re looking at the world around you, but when media is streamed toward your eyes from a projector or display panels built into glasses, it passes through the bump on each contact and comes into focus just in front of the eye." | 14:10 |
| kanzure | ha "Mate choice for cognitive traits: a review of the evidence in nonhuman vertebrates" | 14:12 |
| kanzure | "Song repertoire size in male song sparrows correlates with detour reaching, but not with other cognitive measures" | 14:12 |
| kanzure | dog brain stuff http://diyhpl.us/~bryan/papers2/neuro/The%20world%20from%20a%20dog's%20point%20of%20view:%20a%20review%20and%20synthesis%20of%20dog%20cognition%20research.pdf | 14:15 |
| kanzure | "Does song complexity correlate with problem-solving performance in flocks of zebra finches?" http://lalandlab.st-andrews.ac.uk/documents/Publication207.pdf | 14:18 |
| jrayhawk | ha ha you posted a file:// url | 14:28 |
| kanzure | i blame researchgate | 14:28 |
| kanzure | and myself for updating my browser :( | 14:28 |
| jrayhawk | NEVER AGAIN | 14:29 |
| kanzure | on the internet nobody knows you are a dog, even yourself | 14:37 |
| -!- Guest1309 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has quit [Quit: My Mac has gone to sleep. ZZZzzz…] | 14:38 | |
| kanzure | "Theoretical limits to cognitive control" | 14:40 |
| -!- HEx1 [~HEx@hexwab.plus.com] has quit [Ping timeout: 250 seconds] | 14:41 | |
| fenn | nerp derp | 14:54 |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Ping timeout: 264 seconds] | 15:18 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has joined ##hplusroadmap | 15:21 | |
| -!- chris_99 [~chris_99@unaffiliated/chris-99/x-3062929] has quit [Quit: Ex-Chat] | 15:32 | |
| -!- ebowden [~ebowden@147.69.179.205] has joined ##hplusroadmap | 15:36 | |
| kanzure | that's not the sound a dog makes | 15:40 |
| fenn | in a way i'm glad the dual-n-back intelligence boosting couldn't be repeated; it's bad enough already having to do physical exercise, having to do mental exercise on top of that would waste even more time | 15:49 |
| fenn | and effort | 15:50 |
| fenn | dual-n-back is hard! | 15:50 |
| kanzure | "obviously you should be doing mental exercises while lifting instead of zoning out" | 15:50 |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 252 seconds] | 15:53 | |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 15:53 | |
| fenn | "obviously you should just hire a dominatrix to whip your balls with a branding iron" | 15:53 |
| kanzure | "deciding how to manipulate working memory and perform processing" http://psych.colorado.edu/~oreilly/papers/OReillyFrank06_pbwm.pdf | 15:54 |
| fenn | let me know if you read any of these papers | 15:55 |
| kanzure | i read the ones i can access, although i throw out most of the junk of those | 15:56 |
| kanzure | and the ones that i can't access obviously i can't read | 15:56 |
| kanzure | randall oreilly is a good person that writes good things worth reading | 15:56 |
| kanzure | i hate the phenomena where good papers get cited by everyone even if they misinterpret the results or meaning | 15:59 |
| kanzure | "How [some good paper] reinforces my beliefs about completely unrelated subject x" | 16:00 |
| fenn | would you rather bad papers get cited?~ | 16:09 |
| kanzure | yes they can stay in their own little world or bubble | 16:09 |
| -!- delinquentme [~dingo@74.61.157.78] has quit [Quit: Leaving] | 16:09 | |
| fenn | so how is onetab working for ya | 16:09 |
| -!- delinquentme [~dingo@74.61.157.78] has joined ##hplusroadmap | 16:09 | |
| -!- delinquentme [~dingo@74.61.157.78] has quit [Read error: Connection reset by peer] | 16:10 | |
| kanzure | not using it | 16:11 |
| kanzure | nsh is the culprit here | 16:11 |
| fenn | did he haxxorz you viciously | 16:11 |
| fenn | i heard he was an internet supervillain | 16:12 |
| kanzure | jrayhawk claimed that you were informed about what happened | 16:12 |
| fenn | that you lost your tabs? | 16:13 |
| kanzure | is your comment about internet supervillainary about the downtime? | 16:13 |
| fenn | no | 16:13 |
| kanzure | having lots of tabs does not sound very villainous | 16:13 |
| kanzure | his link to one-tab claims he was running 1506 tabs | 16:14 |
| fenn | that's kinda funny | 16:14 |
| fenn | but also totally relatable | 16:14 |
| fenn | .wik tab guilt | 16:15 |
| nsh | no, it flattens the sessions | 16:15 |
| yoleaux | "Guilt Machine is a musical project by Dutch musician Arjen Anthony Lucassen, owner of other musical projects such as Star One, Ayreon and Ambeon. The project's debut release, On This Perfect Day, was released in August 2009." — http://en.wikipedia.org/wiki/Guilt_Machine | 16:15 |
| fenn | bad robot! | 16:15 |
| kanzure | apparently the prefrontal cortex is our abstraction gland http://www.christofflab.ca/pdfs/Christoff_2009_BrainResearch.pdf | 16:15 |
| kanzure | nsh: can you share link in public? | 16:16 |
| nsh | ctrl-f "restore-all" --> >100 | 16:16 |
| nsh | mm | 16:17 |
| nsh | firefox crashed while i was removing all the FB and Gmail tabs. there's probably a moral in there somewhere | 16:19 |
| kanzure | nsh and i were discussing ways of correctly caching tabs for later reuse | 16:21 |
| nsh | (not that it particularly matters) | 16:21 |
| kanzure | screenshots don't work | 16:22 |
| nsh | ya, yuck | 16:22 |
| kanzure | and you want to capture images, js, css, swf, random remote ajaxy requests to get server data, cookie-setting, etc. | 16:22 |
| nsh | no easy way to share with session sets and save-dates intact, but here: http://www.one-tab.com/page/xCXxOpOITb6avH_Y5x9HIA | 16:22 |
| kanzure | todo: fetus connectome instead of adult connectome. seems simpler. | 16:24 |
| superkuh | I miss Opera session files. | 16:24 |
| superkuh | They were easy enough to grep through, good meta-data. | 16:25 |
| superkuh | Easy to share. | 16:25 |
| nsh | i mooted tabstate as git repo earlier | 16:25 |
| fenn | surely someone's got a export-tabs-to-json plugin | 16:25 |
| kanzure | the reason why i keep tabs open is not because i don't know how to bookmark | 16:25 |
| kanzure | tabs2json is just bookmarking | 16:26 |
| fenn | not if it includes page content | 16:26 |
| fenn | bookmark is usually just url and timestamp (and favicon) | 16:27 |
| superkuh | With Opera sessions it had per-tab history too. | 16:27 |
| superkuh | You could open another person's session file, go to a tab, and see how they arrived there. | 16:27 |
| fenn | i'd like to open my own session file and see what data was on the screen before the website went away forever | 16:28 |
| ebowden | paperbot: http://www.nature.com/neuro/journal/v11/n1/abs/nn2024.html | 16:28 |
| paperbot | http://libgen.org/scimag/get.php?doi=10.1038%2Fnn2024 | 16:28 |
| kanzure | i consider an open tab an attempt at hedging the expected disappearance of the page in the near future | 16:28 |
| kanzure | usually because of high-traffic to the target iste | 16:28 |
| kanzure | and because bookmarking requires actually looking at the content to make a better decision | 16:29 |
| fenn | except not really | 16:30 |
| superkuh | Bookmarking requires figuring out *one* place it fits in the folder hierarchy. | 16:30 |
| fenn | or at least remembering canonical tags | 16:30 |
| superkuh | At least in Firefox. Or you can put it in multiple places by appending nonsense anchors "#dafdasf" to make it a unique URL. | 16:31 |
| -!- Russell [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 16:31 | |
| -!- Russell is now known as Guest47093 | 16:31 | |
| kanzure | Goal-driven cognition in the brain: a computational framework (randall oreilly) http://arxiv.org/pdf/1404.7591.pdf | 16:32 |
| kanzure | Guest47093: hi | 16:32 |
| -!- Guest47093 is now known as russell0 | 16:32 | |
| -!- russell0 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has quit [Client Quit] | 16:34 | |
| kanzure | nice, i like the trend of computational neuroscience papers showing up on arxiv | 16:36 |
| kanzure | "There is also evidence of a common cortical algorithm from surgical manipulation. Functional interchangeability has been demonstrated by an experiment in which visual input was surgically rerouted to auditory cortex in neonatal ferrets, and the mature animals were able to respond to visual stimuli, with retinotopic maps and typical visual receptive fields having developed within their auditory cortex (von Melchner et al., 2000). " | 16:38 |
| kanzure | last quote is from "Towards a universal cortical algorithm: Examining hierarchical temporal memory in light of frontal cortical function" http://arxiv.org/pdf/1411.4702.pdf | 16:39 |
| kanzure | "Lennie (2003) found that given its energy usage constraints, less than 2% of the brain's neurons could afford to be significantly active at any given time." well okay, but what counts as active.. | 16:44 |
| fenn | Scientists prove you only use 2% of your brain! | 16:48 |
| kanzure | "Ferrier (1874) first mapped the motor areas using direct electrical stimulation. He found that stimulation of specific areas of motor cortex would produce movement in corresponding muscle groups, resulting in a rough somatotopic map. Brief stimulation would produce muscle twitches, while more prolonged stimulation would produce coordinated sequences of seemingly meaningful movements, such as stepping or reaching." | 16:49 |
| kanzure | 1874.. hrm. | 16:50 |
| kanzure | the failure modes described in the "Functions of the prefrontal cortex" section are pretty interesting | 16:52 |
| kanzure | hmm nevermind. this is just a laundry list of things that brains do. | 16:55 |
| fenn | "Suddenly, I was immersed in the brightest light I have ever seen. I moved my head from side to side and all I could see was bright white light, similar in appearance to arc welding light. This next part is strange. I distinctly remember feeling weightless, and that my feet were no longer touching the ground. For some reason, it felt like a number of seconds transpired, even though I | 16:57 |
| fenn | realize that lightning strikes are instantaneous. A deafening explosion followed, and I found myself on the ground with the others. Sean was collapsed and huddled on his knees. Smoke was pouring from his back. I rushed over to him and checked his pulse and breathing. He was still alive. I put out the embers on his back and elbows and carried him down the path towards the parking lot, | 16:57 |
| fenn | with the rest of the group following. | 16:57 |
| kanzure | huh page 56 describes where exactly which neurons and projections are effected by reinforcement learning | 17:07 |
| nmz787 | kanzure: any ideas? http://paste.pound-python.org/show/TfYN7sFEjBTgTGTOzjoo/ | 17:10 |
| nmz787 | forgot to add there that I am using ndk-r10c on Ubuntu 14.04 and tried with opencv 2.4.9 and 2.4.10 (and the respective versions of the opencv android download) | 17:11 |
| -!- Boscop_ [me@46.246.78.102] has joined ##hplusroadmap | 17:14 | |
| fenn | nmz787: http://answers.ros.org/question/191380/opencv-surf-is-there-but-cant-be-found/ "If you don't have the symbol, you need to add something to your link line." | 17:15 |
| nmz787 | but all I am doing is calling some program called ndk-build... I don't know where linker options even would go | 17:16 |
| * nmz787 looks at that link | 17:16 | |
| fenn | you have a line there that says LOCAL_LDLIBS += -llog | 17:17 |
| -!- Boscop [~me@unaffiliated/boscop] has quit [Ping timeout: 240 seconds] | 17:17 | |
| fenn | i dont know what provides vtable tho | 17:17 |
| -!- snuffeluffegus [~snuff@ps357888.dreamhost.com] has quit [Quit: Leaving] | 17:18 | |
| fenn | maybe add LOCAL_LDLIBS += -llibopencv_nonfree -llibopencv_features2d | 17:18 |
| nmz787 | reading this now | 17:19 |
| nmz787 | http://docs.opencv.org/doc/tutorials/introduction/android_binary_package/android_dev_intro.html#ndk-build-cli | 17:19 |
| fenn | to Android.mk | 17:19 |
| nmz787 | the weird thing is building the shared libnonfree.so didn't give errors, but was much smaller than the pre-compiled found on some blog | 17:20 |
| fenn | compilation options could be different | 17:21 |
| fenn | optimization often adds to disk size in exchange for speed | 17:22 |
| nmz787 | mm | 17:22 |
| nmz787 | and I am using a newer version of the library too | 17:22 |
| fenn | did you try adding my line to the makefile? | 17:22 |
| nmz787 | yeah here is the result http://paste.pound-python.org/show/GlW157fFZhYfyqJ0oHdz/ | 17:24 |
| nmz787 | still same | 17:24 |
| nmz787 | but i used the rebuild all and verbose options for ndk-build | 17:24 |
| nmz787 | it mentions something about your additions at the top | 17:25 |
| fenn | it can't find the libraries | 17:25 |
| nmz787 | well the blog said to copy them to the source directory after building them | 17:26 |
| nmz787 | which i did | 17:26 |
| fenn | oh, the error is occurring while building features2d so you wouldn't have that library yet | 17:27 |
| nmz787 | so the error is silent? | 17:27 |
| fenn | no, i'm just mixed up | 17:27 |
| nmz787 | I added your changes to the app's Android.mk | 17:27 |
| nmz787 | not the libnonfree build's Android.mk | 17:27 |
| fenn | the error is undefined reference to 'vtable for cv::SIFT' | 17:27 |
| nmz787 | oh, this is the blog https://sites.google.com/site/wghsite/technical-notes/sift_surf_opencv_android | 17:28 |
| nmz787 | yeah | 17:28 |
| nmz787 | when I use his pre-compiled ones it works | 17:28 |
| fenn | precompiled whats | 17:29 |
| nmz787 | libnonfree.so | 17:29 |
| fenn | you have two makefiles, Android.mk and Application.mk, yes? | 17:31 |
| nmz787 | 2 sets of those | 17:32 |
| fenn | why 2 sets | 17:32 |
| nmz787 | one set for building the opencv library, and one for the program that is trying to use those libraries | 17:32 |
| nmz787 | to do some image detection and drawing features | 17:32 |
| fenn | -_- | 17:32 |
| fenn | this looks sketchy: LOCAL_SRC_FILES := libnonfree.so | 17:34 |
| nmz787 | why? i copied it into the source dir | 17:36 |
| fenn | it's not a source file | 17:36 |
| fenn | maybe "it works" but i have no idea what it's doing | 17:36 |
| nmz787 | ah | 17:38 |
| nmz787 | I tried copying the .so file to the opencv dir with all the other .so files | 17:38 |
| nmz787 | didn't work | 17:39 |
| fenn | is the app supposed to even work on your PC? | 17:41 |
| fenn | "Use "adb push" tools to upload test_sift, libnonfree.so, libopencv_java.so along with a test image to your device" | 17:42 |
| kanzure | adb push works fine with emulated android devices | 17:43 |
| fenn | maybe he hasn't gotten that far yet | 17:44 |
| fenn | debugging cross-compilation over IRC, yay! | 17:45 |
| kanzure | https://github.com/pdiasalmeida/openCVAndroid-ndkSample | 17:46 |
| kanzure | downside is that this person uses eclipse so brace yourself for awfulness | 17:47 |
| -!- CheckDavid [uid14990@gateway/web/irccloud.com/x-udgtfhtywffecsup] has quit [Quit: Connection closed for inactivity] | 17:47 | |
| fenn | i once tried to use eclipse but my computer couldn't handle it | 17:47 |
| kanzure | you'll have to wait until java bytecode can be JIT converted to thalamocortical inhibitory magic | 17:48 |
| fenn | huh | 17:49 |
| fenn | or i could use gcc and vim | 17:49 |
| fenn | gcj | 17:50 |
| fenn | or i could hire a dominatrix to whip my balls with a branding iron | 17:51 |
| kanzure | i am reading http://psych.colorado.edu/~oreilly/papers/oreillyhazyherdip.pdf so it leaked through (it describes randall oreilly's biologically-accurate simulator, leabra) | 17:51 |
| nmz787 | well I found some info on building the complete android opencv libraries, rather than using the precompiled libraries and only compiling the nonfree library myself... I had already compiled opencv myself and have been using it, and the android stuff is compiling now | 17:54 |
| nmz787 | the android libs | 17:54 |
| nmz787 | so maybe I can make install when it's done, and remove that local_src reference | 17:55 |
| nmz787 | in the app's build .mk | 17:55 |
| fenn | you still need that because it references the non-free shared library you built in the previous step | 17:55 |
| fenn | i think all it does is copy it to ibs/armeabi/libopencv_java.so | 17:55 |
| fenn | libs* | 17:55 |
| nmz787 | well that's what I'm saying, I'm not /just/ building the nonfree shared lib... i'm building all opencv for android libs (inclduding nonfree) | 17:56 |
| nmz787 | so if anything I expect it to work, or crap out even more later | 17:57 |
| nmz787 | when I try to build the app | 17:57 |
| nmz787 | i didn't even end up geting that $40 tablet, went back today and it was $50 and said screw it | 17:57 |
| fenn | bah | 17:57 |
| fenn | it's still cheap at $50 | 17:57 |
| kanzure | you can probably get it for $10 on alibaba | 17:58 |
| nmz787 | it would be pretty sweet to control the FIB with my phone though | 17:58 |
| nmz787 | i was looking at the arduino USB OTG driver for android | 17:58 |
| nmz787 | there was a nice make mag writeup with an app | 17:58 |
| nmz787 | http://makezine.com/projects/make-36-boards/android-arduino-led-strip-lights/ | 17:58 |
| * fenn prepares to be underwhelmed | 17:59 | |
| fenn | "we fried a phone" really? really? | 18:00 |
| kanzure | "The glass brain visualization on the upper left projects simulated neural activity into the anatomical locations of simulated brain areas, for easier direct comparison with neuroimaging and other data." | 18:00 |
| kanzure | was this really disputed? "Thus, we reject the notion of a neural code that posits meaning in individual neural signals, and accept the consequence that it is not possible for neurons to rapidly change what they encode — that would just confuse the other neurons (O’Reilly, 2010). Instead, neural representations must be relatively stable over time, to enable a given receiving neuron to properly learn the statistics of the patterns of ... | 18:00 |
| kanzure | ... activity over its inputs." | 18:00 |
| fenn | presumably the phone doesn't power the LEDs so why not run the arduino on that power supply also | 18:02 |
| kanzure | i could maybe accept something about neural ensemble code stuff near inputs, at most | 18:02 |
| fenn | was MAKE always this bad? | 18:04 |
| kanzure | i hear that the first five or so issues don't tend to induce cancer as much | 18:06 |
| fenn | nmz787: i'd recommend ignoring anything those people have to say as they are clearly a) irresponsible b) don't know what they are doing and c) write bad code | 18:08 |
| nmz787 | lol | 18:08 |
| nmz787 | it works | 18:08 |
| nmz787 | good enough to get me started | 18:08 |
| nmz787 | maybe | 18:08 |
| fenn | until you plug in the wrong phone and it burns out the usb port | 18:08 |
| nmz787 | nah | 18:08 |
| fenn | nah man, NAH! | 18:08 |
| nmz787 | that seemed to be their fault for using a non-otg device | 18:08 |
| fenn | it's their fault for doing it wrong | 18:09 |
| kanzure | https://lucien0maverick.files.wordpress.com/2014/07/squidward1.jpg | 18:11 |
| kanzure | "Moving down a level in the hierarchy, the equivalent of posterior cortex in the affective domain is the basolateral amygdala (BLA), which is anatomically at the same level as the hippocampus in what is known as the “archicortex” or ancient cortex. The BLA is densely interconnected with the OFC and the ACC, and it is known to encode both US’s and CS’s." | 18:11 |
| kanzure | death by acronyms | 18:11 |
| nmz787 | fenn: ok so how do I find all the opencv ldlib inlcludes I need? is that what pkgconfig does? | 18:15 |
| fenn | this sounds like some awful redhat thing | 18:17 |
| nmz787 | (opencv compiled successfully with nonfree as part of it, so now I shouldn't have to reference any local modules, I would assume I should be able to just reference the opencv libs dir at once, since it has them all) | 18:17 |
| fenn | sure, okay | 18:17 |
| nmz787 | now it's just complaining about all the opencv calls :P | 18:18 |
| fenn | obviously it can't find the libraries | 18:18 |
| nmz787 | how did you come up with -llibopencv_features2d | 18:19 |
| fenn | i made it up | 18:20 |
| nmz787 | how can i find the rest with that syntax? | 18:20 |
| kanzure | ld | 18:20 |
| kanzure | ldd | 18:20 |
| nmz787 | but where does the path get specified? | 18:20 |
| kanzure | LD_SOMETHING_PATH | 18:20 |
| kanzure | probably LD_LIBRARY_PATH | 18:20 |
| kanzure | or LD_LOAD_PATH | 18:20 |
| nmz787 | oh, I have to export that? | 18:21 |
| fenn | this is android crap | 18:21 |
| kanzure | android linux probably has a few default places to find shared libraries | 18:21 |
| kanzure | "Surely you must all be joking: An outsider's critique of quantum physics" http://arxiv.org/pdf/1109.0880.pdf (same author) | 18:23 |
| kanzure | nsh: ^ | 18:23 |
| nsh | cool | 18:23 |
| -!- cpopell [~cpopell@c-76-26-144-132.hsd1.dc.comcast.net] has quit [Ping timeout: 258 seconds] | 18:24 | |
| -!- cpopell [~cpopell@c-76-26-144-132.hsd1.dc.comcast.net] has joined ##hplusroadmap | 18:24 | |
| fenn | so in ndk-build there is LOCAL_LDLIBS LOCAL_STATIC_LIBRARIES LOCAL_SHARED_LIBRARIES | 18:25 |
| fenn | in normal makefiles you should be able to add directories with LDFLAGS := -R$(your_lib_dir) | 18:27 |
| kanzure | "The cortical connectivity of the prefrontal cortex in the monkey brain" http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3161133/ | 18:36 |
| kanzure | so uh what did we need the connectome for, again? | 18:36 |
| kanzure | "Monkey to human comparative anatomy of the frontal lobe association tracts" http://cogimage.dsi.cnrs.fr/hmtc/references/files/pubs_hmtc/Thiebaut%20de%20SchottenMonkey%20to%20human%20comparative%20anatomyCortex2011-1.pdf | 18:37 |
| -!- sheena [~home@S0106c8be196316d1.ok.shawcable.net] has joined ##hplusroadmap | 18:53 | |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 272 seconds] | 18:53 | |
| -!- sheena [~home@S0106c8be196316d1.ok.shawcable.net] has quit [Read error: Connection reset by peer] | 18:53 | |
| kanzure | haha robert freitas is a bitcoiner | 18:54 |
| kanzure | but lost it all on mtgox :) | 18:54 |
| fenn | ouch | 18:57 |
| fenn | my touchpad stopped doing tap or scroll, and my keyboard auto-repeat seems slower (longer delay) any ideas why this would happen or where to look? | 18:58 |
| kanzure | persistent issue happen many times? | 18:58 |
| fenn | no | 18:58 |
| kanzure | start by rmmod and modprobe psmouse? | 18:59 |
| fenn | hey that worked | 18:59 |
| kanzure | keyboard issues may be separate | 18:59 |
| -!- Russell [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 19:03 | |
| fenn | since "upgrading" to saucy i've been having issues with mouse and keyboard focus not working as expected, and wmi_event suspend and power buttons not doing the things they should | 19:03 |
| -!- Russell is now known as Guest50322 | 19:03 | |
| kanzure | Guest50322: howdy again | 19:03 |
| -!- Guest50322 is now known as russell0 | 19:05 | |
| russell0 | hey bryan :) | 19:06 |
| kanzure | russell0: i am reading these things, http://diyhpl.us/~bryan/papers2/neuro/randall-oreilly/ | 19:13 |
| kanzure | there seems to be lots of progress sans complete connectomes | 19:15 |
| kanzure | like comparative neuroanatomy of monkey/human prefrontal cortex http://cogimage.dsi.cnrs.fr/hmtc/references/files/pubs_hmtc/Thiebaut%20de%20SchottenMonkey%20to%20human%20comparative%20anatomyCortex2011-1.pdf and http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3161133/ | 19:15 |
| russell0 | thanks for the randall oreilly papers | 19:17 |
| kanzure | i want to make a trap for people that use too many abbreviations | 19:33 |
| -!- russell0 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has quit [Quit: My Mac has gone to sleep. ZZZzzz…] | 19:51 | |
| -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 20:05 | |
| -!- Russell [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 20:22 | |
| -!- Russell is now known as Guest93882 | 20:22 | |
| -!- Guest93882 is now known as russell0 | 20:23 | |
| -!- russell0 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has left ##hplusroadmap [] | 20:26 | |
| -!- russell0 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 20:28 | |
| kanzure | http://greymattersjournal.com/wp-content/uploads/2013/12/orca-human-brain-comparison.gif | 20:37 |
| -!- Beatzebub_ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 21:05 | |
| nmz787 | paperbot: http://link.springer.com/chapter/10.1007/978-3-7091-0706-5_4 | 21:07 |
| -!- Beatzebub [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Ping timeout: 244 seconds] | 21:09 | |
| nmz787 | http://web.archive.org/web/20080514131008/http://www.space-mining.com/GMI-FINAL.htm | 21:12 |
| nmz787 | .title | 21:12 |
| yoleaux | PROPOSAL # 04-X1 | 21:12 |
| nmz787 | .wik mond process | 21:13 |
| yoleaux | "The Mond process, sometimes known as the carbonyl process is a technique created by Ludwig Mond in 1890 to extract and purify nickel. The process was used commercially before the end of the 19th century. This process converts nickel oxides into pure nickel." — http://en.wikipedia.org/wiki/Mond_process | 21:13 |
| nmz787 | "This process makes use of the fact that carbon monoxide complexes with nickel readily and reversibly to give nickel carbonyl. No other element forms a carbonyl compound under the mild conditions used in the process." | 21:13 |
| -!- Viper168_ [~Viper@unaffiliated/viper168] has joined ##hplusroadmap | 21:23 | |
| -!- Boscop_ [me@46.246.78.102] has quit [Ping timeout: 264 seconds] | 21:23 | |
| -!- Viper168 [~Viper@unaffiliated/viper168] has quit [Ping timeout: 240 seconds] | 21:24 | |
| -!- rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 21:28 | |
| -!- rayston_ [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 21:29 | |
| -!- rayston_ [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Read error: Connection reset by peer] | 21:29 | |
| -!- rayston is now known as Rayston | 21:30 | |
| -!- yashgaroth [~ffffff@2606:6000:cb85:6a00:8c51:9d75:8b6e:768] has quit [Quit: Leaving] | 21:32 | |
| -!- Beatzebub__ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 22:05 | |
| nmz787 | so it seems that I was able to link successfully after compiling opencv-android completely from scratch | 22:06 |
| nmz787 | but when I try to run the program in adb shell, it says permission denied | 22:06 |
| nmz787 | last night I was using eclipse (with half-working code) and it was able to run the 'activity' example project app | 22:07 |
| nmz787 | it would be nice to make the apk from the command line | 22:07 |
| -!- Beatzebub_ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Ping timeout: 255 seconds] | 22:09 | |
| -!- ebowden [~ebowden@147.69.179.205] has quit [Remote host closed the connection] | 22:23 | |
| russell0 | going around... https://www.facebook.com/CicretAppandBracelet | 22:27 |
| russell0 | wrist-mounted cell phone projector and interface | 22:27 |
| russell0 | "smart watch" | 22:27 |
| -!- ebowden [~ebowden@147.69.179.205] has joined ##hplusroadmap | 22:28 | |
| -!- pete4242 [~smuxi@boole.london.hackspace.org.uk] has quit [Read error: Connection reset by peer] | 22:56 | |
| -!- Beatzebub_ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has joined ##hplusroadmap | 23:05 | |
| nmz787 | whoo, finally got it | 23:09 |
| -!- Beatzebub__ [~beatzebub@S0106b81619e8ecee.gv.shawcable.net] has quit [Ping timeout: 255 seconds] | 23:09 | |
| -!- lichen [~lichen@c-50-139-11-6.hsd1.or.comcast.net] has quit [Quit: Lost terminal] | 23:10 | |
| fenn | you learned how to use chmod successfully? | 23:16 |
| fenn | how long does SIFT take on your android device? | 23:16 |
| -!- russell0 [~textual@cpe-74-73-107-82.nyc.res.rr.com] has quit [Quit: Textual IRC Client: www.textualapp.com] | 23:18 | |
| nmz787 | it took longer than I expected, but not terribly longer than my 4 or 5 year old single-core android 2.3 device | 23:18 |
| fenn | um, so, seconds? milliseconds? | 23:18 |
| -!- Russell [~textual@cpe-74-73-107-82.nyc.res.rr.com] has joined ##hplusroadmap | 23:18 | |
| nmz787 | but my laptop can do sift very very quickly, so I think there are some settings to be tweaked | 23:19 |
| nmz787 | seconds | 23:19 |
| nmz787 | like 10 | 23:19 |
| -!- Russell is now known as Guest78474 | 23:19 | |
| nmz787 | wait | 23:20 |
| nmz787 | let me try again | 23:20 |
| nmz787 | that time seemed more like a minute | 23:21 |
| nmz787 | ok started fresh, and there are timestamps on the logs | 23:22 |
| -!- Rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Ping timeout: 272 seconds] | 23:22 | |
| nmz787 | fenn: http://paste.pound-python.org/show/UsypszmIozx5lQ4WVWkX/ | 23:22 |
| nmz787 | so I guess 44 seconds | 23:23 |
| nmz787 | but it definitely seemed much faster to find keypoints between two images and print their offset, like a few FPS | 23:24 |
| nmz787 | this is the run_demo c++ function http://paste.pound-python.org/show/bg5vDq0cU3WIzBfB1xnB/ | 23:26 |
| nmz787 | let me find the example that comes with opencv | 23:26 |
| fenn | only needs a few keypoints to do that | 23:26 |
| fenn | the real beef is in the GPU or whatever it's called in a mobile processor | 23:27 |
| fenn | i know people have done GPU SIFT on a mobile device, but how... | 23:27 |
| fenn | too many layers of semi-intentional obfuscation | 23:28 |
| nmz787 | I wonder if I was using SURF then | 23:30 |
| -!- Rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 23:30 | |
| nmz787 | yeah I don't think I compiled GPU opencv, if that exists for my phone | 23:30 |
| nmz787 | there is some way to change what instruction set it uses, which may support higher performance | 23:31 |
| -!- Rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has quit [Excess Flood] | 23:32 | |
| -!- Rayston [~rayston@ip68-106-242-42.ph.ph.cox.net] has joined ##hplusroadmap | 23:35 | |
| fenn | hmm i may have misunderstood.. it was from the oxford active vision group, they had some papers like "Simultaneous 3D Tracking and Reconstruction on a Mobile Phone" and "Fast Feature Detection with a Graphics Processing Unit Implementation" | 23:37 |
| fenn | but these are different projects | 23:38 |
| -!- Darius__ [~quassel@71-94-157-171.static.mtpk.ca.charter.com] has joined ##hplusroadmap | 23:40 | |
| fenn | .title http://aut.researchgateway.ac.nz/handle/10292/7991 | 23:41 |
| yoleaux | GPU accelerated feature algorithms for mobile devices | 23:41 |
| nmz787 | anyway I was using the demo from the opencv samples directory | 23:41 |
| nmz787 | whichever one used the baboon image | 23:42 |
| * nmz787 sleeps | 23:42 | |
| -!- Darius__ [~quassel@71-94-157-171.static.mtpk.ca.charter.com] has quit [Ping timeout: 240 seconds] | 23:45 | |
Generated by irclog2html.py 2.15.0.dev0 by Marius Gedminas - find it at mg.pov.lt!
