I thought it was obvious the point I was making, but I guess I wasn't clear. When you live long enough to *personally* see the result of abusing the environment, you're more likely to do something about it.
A lot of science fiction has explored life extension.. of course, a lot of people don't read, and a lot less read science fiction, so I expect that we'll rehash everything as the technology becomes available.
Does make you wonder what they do to citizens who just say no to an incoming search. They can't refuse them entry.. They have to arrest them or let them go. If USians hadn't turned into such pussies since 9/11, *someone* would have gotten themselves arrested by now and tested the law, but no, USians are afraid of being arrested now.
As much as I understand the point you are trying to make, I really don't think it is as big of a practical issue as much as it is a psychological one. All the people who contribute to the Linux kernel know maybe 1/1000th of the entire code base, because that's all they need to know. Same goes for Open Office or any other large code base. I remember fixing something in Open Office a few years ago.. I looked at the text on the dialog, searched the code, didn't find the text. That stumped me for about an hour, until I figured out that they were internationalization fans.. I found the string I was looking for in a language file. Next to it was an identifier.. I searched for that in the code, found the part which was responsible for the dialog. If I remember correctly, it was some kind of domain specific language for dialogs. So I studied that for a while and found the identifier for the function that was called when the button I was interested in was pressed. That function was written in C++.. it's probably written in Java now. About 20 seconds after looking at the function I saw the obvious error, and fixed it. I sent the patch to the list, they said thanks, ignored the patch and just fixed it themselves. I implemented a couple of other hacks that I wanted to the Presenter app, posted em to the list, they were picked up in some later releases.
This is what programming in the large is all about. Most software is worked on by a team of developers, few of whom understand the entire code base. You need to use "beacons" to navigate around the code and find the local relevant portion that corresponds to the dynamic behavior. This is classic software maintenance.. see Brooks, 1983. I, personally, think the standard mechanism is not good enough.. There should be better tool support for matching program behavior to the source code, and programmers should learn to use them. I wrote a function tracing tool for this reason. Although, looking at that page now, I see this isn't the version that lets me attach to an already running pid. I should really update that. Can't remember if it does multiple threads either.. hmm.
Picking up a new code-base quickly is a skill that many people in the Open Source community have.. and you will never learn it if you're just defeatist about it.
Yes, you're right. The thing that makes it patentable or not is the method. If it is trivial to combine X with Y, then it is trivial and may not receive patent protection. If, however, you need to invent something to combine X with Y, as you most always do, then it likely is patentable.
TRBMs have been used in DBNs, look at Learning Multilevel Distributed Representations for High-Dimensional Sequences, Ilya Sutskever and Geoffrey Hinton, AISTATS 2007. But yeah, if you're looking for a job, Numenta are a good place to look. Of course, once you join the company you'll think Numenta's technology is the "one true path" and not even bother looking at the rest of the field;)
All I want is some proof. Is that so hard? No, and you're right to ask for it. Skepticism is healthy.
There's three main components to OpenCog:
* The architecture
* The Probabilistic Logic Network algorithm
* The competent generic algorithms system (MOSES)
and it is openly admitted that this isn't enough.. some sub-symbolic algorithms will be needed to complement all this "neat" (although probabilistically weighted) work.
And then hanging on the side is all the Natural Language Processing stuff, and the embodied simulation stuff, both of which are meant to generate interesting input for OpenCog to process.
To date, the PLN algorithm hasn't not been made public, although that could change in the next month. The MOSES stuff is public, but not integrated yet. The Natural Language Processing stuff is written in Java (the rest of the framework is written in C++) and this won't change as the XML interfaces between the two are sufficiently powerful that there is just no need.
Probably not until September 2008 or later will all these parts start coming together.. and it'll likely be well into next year before something "impressive" can be demoed without outright cheating. Then, hopefully, things will take off. But who knows. We have to wait and see and optimistic predictions do more harm than good.
I think the Deep Belief Networks of Hinton et al are way ahead of Numenta.. in that they are real science with measurable results that has been reproduced by multiple implementations. The 2006 paper that started it all and Hinton's presentation on google video:
That's not the way prior art works. Patents cover methods, not ideas. So a method that applies an old idea to a new situation can indeed be patentable. If you work for IBM, or some other company that has a bonus scheme for patent filing, one way to come up with shit to get patented is simply to make a list of all the new technologies out there and figure out how to apply old ideas using them.
There's reporting "labeling is borked!!" and there is reporting "with configuration X, running version Y, I had problem T and this prevented me from doing real work".. but yeah, if no-one is interested in fixing the bug, then no-one is interested in fixing the bug.. It doesn't matter if Sun is backing the project or not. If they don't consider it a priority then its up to you to fix it. Step up. If you can't code, post on the mailing list that you're willing to help anyone who can to reproduce the issue.. even if that means sending them any special hardware that might be required.
Slashdot Mirror
"Everything is pretty much run by [engineering] -- PMs and testers are conspicuously absent from the process."
Oh what a fucking nightmare!
Check out the Linux kernel some time.
http://video.google.com/url?docid=3414999620283973916
Command and control that a military would envy.
If you RTFA you'd see that the thing that is spinning, is a hologram.
And who makes the laws?
I thought it was obvious the point I was making, but I guess I wasn't clear. When you live long enough to *personally* see the result of abusing the environment, you're more likely to do something about it.
No, they're dumb. Fox News was unable to find people who could act dumb, so now they just hire people who really, (really, really) are dumb.
Please warn us when linking to Fox News. Jesus those people are dumb.
A lot of science fiction has explored life extension.. of course, a lot of people don't read, and a lot less read science fiction, so I expect that we'll rehash everything as the technology becomes available.
People don't give a shit about the planet because they know they will be dead long before it is.
Give them eternal life and watch how quickly they become militant greenies.
Say a particular 4 letter lobbying organization was offering these ISPs money to curb P2P usage.. would that be legal?
Kinda sounds like tortuous interference to me.
Does make you wonder what they do to citizens who just say no to an incoming search. They can't refuse them entry.. They have to arrest them or let them go. If USians hadn't turned into such pussies since 9/11, *someone* would have gotten themselves arrested by now and tested the law, but no, USians are afraid of being arrested now.
As much as I understand the point you are trying to make, I really don't think it is as big of a practical issue as much as it is a psychological one. All the people who contribute to the Linux kernel know maybe 1/1000th of the entire code base, because that's all they need to know. Same goes for Open Office or any other large code base. I remember fixing something in Open Office a few years ago.. I looked at the text on the dialog, searched the code, didn't find the text. That stumped me for about an hour, until I figured out that they were internationalization fans.. I found the string I was looking for in a language file. Next to it was an identifier.. I searched for that in the code, found the part which was responsible for the dialog. If I remember correctly, it was some kind of domain specific language for dialogs. So I studied that for a while and found the identifier for the function that was called when the button I was interested in was pressed. That function was written in C++.. it's probably written in Java now. About 20 seconds after looking at the function I saw the obvious error, and fixed it. I sent the patch to the list, they said thanks, ignored the patch and just fixed it themselves. I implemented a couple of other hacks that I wanted to the Presenter app, posted em to the list, they were picked up in some later releases.
This is what programming in the large is all about. Most software is worked on by a team of developers, few of whom understand the entire code base. You need to use "beacons" to navigate around the code and find the local relevant portion that corresponds to the dynamic behavior. This is classic software maintenance.. see Brooks, 1983. I, personally, think the standard mechanism is not good enough.. There should be better tool support for matching program behavior to the source code, and programmers should learn to use them. I wrote a function tracing tool for this reason. Although, looking at that page now, I see this isn't the version that lets me attach to an already running pid. I should really update that. Can't remember if it does multiple threads either.. hmm.
Dude, speak for yourself.
Picking up a new code-base quickly is a skill that many people in the Open Source community have.. and you will never learn it if you're just defeatist about it.
Yes, you're right. The thing that makes it patentable or not is the method. If it is trivial to combine X with Y, then it is trivial and may not receive patent protection. If, however, you need to invent something to combine X with Y, as you most always do, then it likely is patentable.
They're talking about the language that the programs the genetic algorithms evolve.
Right now, whoever you are.
If you'd like some help getting involved, drop me an email.
TRBMs have been used in DBNs, look at Learning Multilevel Distributed Representations for High-Dimensional Sequences, ;)
Ilya Sutskever and Geoffrey Hinton, AISTATS 2007. But yeah, if you're looking for a job, Numenta are a good place to look. Of course, once you join the company you'll think Numenta's technology is the "one true path" and not even bother looking at the rest of the field
Welcome to the broken patent system.
Why do you think so many things have clocks in them?
There's three main components to OpenCog:
* The architecture
* The Probabilistic Logic Network algorithm
* The competent generic algorithms system (MOSES)
and it is openly admitted that this isn't enough.. some sub-symbolic algorithms will be needed to complement all this "neat" (although probabilistically weighted) work.
And then hanging on the side is all the Natural Language Processing stuff, and the embodied simulation stuff, both of which are meant to generate interesting input for OpenCog to process.
To date, the PLN algorithm hasn't not been made public, although that could change in the next month. The MOSES stuff is public, but not integrated yet. The Natural Language Processing stuff is written in Java (the rest of the framework is written in C++) and this won't change as the XML interfaces between the two are sufficiently powerful that there is just no need.
Probably not until September 2008 or later will all these parts start coming together.. and it'll likely be well into next year before something "impressive" can be demoed without outright cheating. Then, hopefully, things will take off. But who knows. We have to wait and see and optimistic predictions do more harm than good.
I think the Deep Belief Networks of Hinton et al are way ahead of Numenta.. in that they are real science with measurable results that has been reproduced by multiple implementations. The 2006 paper that started it all and Hinton's presentation on google video:
http://www.gatsby.ucl.ac.uk/~ywteh/research/ebm/nc2006.pdf
http://video.google.com.au/videoplay?docid=228784531481853811
A formal analysis:
http://www.cs.utoronto.ca/~ilya/pubs/2007/inf_deep_net_utml.pdf
Application to natural language processing:
http://www.cs.swarthmore.edu/~meeden/cs81/s08/DahlLaTouche.pdf
http://www.machinelearning.org/proceedings/icml2007/papers/425.pdf
Reproducing Hinton and extension to and evaluation in other domains:
http://www.machinelearning.org/proceedings/icml2007/papers/331.pdf
Use in Computer animation of facial expressions:
http://aclab.ca/users/josh/downloads/pubs/23_Susskind_Hinton_Movellan_Anderson.pdf
Most impressive:
http://www.cs.utoronto.ca/~ilya/pubs/2007/aistats_multilayered.pdf
A C++ implementation (although it has much Python love):
http://plearn.berlios.de/
So yeah, there's some pretty good demonstrations of how powerful DBNs are.. Numenta is lagging behind.
http://www.opencog.org/wiki/Main_Page
http://www.agiri.org/OpenCog_AGI-08.pdf
http://justingibbs.com/how-to-make-singularity-bearable-in-its-infancy
http://www.innergybv.biz/blog/?p=175
http://ieet.org/index.php/IEET/more/goertzel20080620/#When:22:49:00Z
http://xlaurent.blogspot.com/2008/06/opensim-for-opencog.html
There's a number of GSoC projects for OpenCog currently underway also:
http://code.google.com/soc/2008/siai/about.html
So the first release should be very interesting.
That's not the way prior art works. Patents cover methods, not ideas. So a method that applies an old idea to a new situation can indeed be patentable. If you work for IBM, or some other company that has a bonus scheme for patent filing, one way to come up with shit to get patented is simply to make a list of all the new technologies out there and figure out how to apply old ideas using them.
They should have called it "Google Mail" from the beginning. I still have people asking me what the hell gmail is.
There's reporting "labeling is borked!!" and there is reporting "with configuration X, running version Y, I had problem T and this prevented me from doing real work".. but yeah, if no-one is interested in fixing the bug, then no-one is interested in fixing the bug.. It doesn't matter if Sun is backing the project or not. If they don't consider it a priority then its up to you to fix it. Step up. If you can't code, post on the mailing list that you're willing to help anyone who can to reproduce the issue.. even if that means sending them any special hardware that might be required.