Comments for OpenCog Brainwave http://blog.opencog.org The latest developments in building an open-source mind Fri, 23 Mar 2012 21:15:00 +0000 hourly 1 http://wordpress.org/?v=3.0.1 Comment on Genetic Crossover in MOSES by Linas Vepstas http://blog.opencog.org/2012/03/20/genetic-crossover-in-moses/#comment-91 Linas Vepstas Fri, 23 Mar 2012 21:15:00 +0000 http://blog.opencog.org/?p=370#comment-91 Ahh, but I didn't want to ...  1+2+3 every 10 uses the simplex extrapolation for 9 steps, and then a full neighborhood rescan every tenth generation.  As you can see, it advances rapidly then stalls.1+2+3 rescan uses simplex, until there is no improvement, and then it does a full neighborhood scan. rev6654 is whatever bzr revision 6654 does. nn16 is bzr rev 6600 but with a "last chance" rescan disabled. nn18 is same as nn16, but with a different random seed. Notice how dramatically different they are. nn28 is bzr rev 6666, i.e. the "final" code. Because of the sharp dependence on the initial random seed (and thus, implicit random seed changes when algos access the RNG in a different order) makes it hard to compare results.  Ideally, one should repeate measurements, averaging over at least 20 or better, 100 different random seeds.   This is hard to do... More details in the text file "diary-performance.txt" in the source code tree.  Ahh, but I didn’t want to … 
1+2+3 every 10 uses the simplex extrapolation for 9 steps, and then a full neighborhood rescan every tenth generation.  As you can see, it advances rapidly then stalls.1+2+3 rescan uses simplex, until there is no improvement, and then it does a full neighborhood scan.

rev6654 is whatever bzr revision 6654 does.

nn16 is bzr rev 6600 but with a “last chance” rescan disabled.

nn18 is same as nn16, but with a different random seed. Notice how dramatically different they are.

nn28 is bzr rev 6666, i.e. the “final” code.

Because of the sharp dependence on the initial random seed (and thus, implicit random seed changes when algos access the RNG in a different order) makes it hard to compare results.  Ideally, one should repeate measurements, averaging over at least 20 or better, 100 different random seeds.   This is hard to do…

More details in the text file “diary-performance.txt” in the source code tree. 

]]> Comment on Genetic Crossover in MOSES by Nil Geisweiller http://blog.opencog.org/2012/03/20/genetic-crossover-in-moses/#comment-90 Nil Geisweiller Wed, 21 Mar 2012 06:47:00 +0000 http://blog.opencog.org/?p=370#comment-90 Awesome work, Linas. Your last remark "the algos which advance the fastest initially seem to have trouble advancing later on" is indeed pretty interesting. Maybe MOSES could do the tweaking itself depending on the target score or number of evals, etc, possibly via some form of metalearning. Could you briefly detail the legend of that last graph? Thanks... Awesome work, Linas.

Your last remark “the algos which advance the fastest initially seem to have trouble advancing later on” is indeed pretty interesting. Maybe MOSES could do the tweaking itself depending on the target score or number of evals, etc, possibly via some form of metalearning.

Could you briefly detail the legend of that last graph?

Thanks…

]]> Comment on Tuning Metalearning in MOSES by chhean http://blog.opencog.org/2012/02/07/tuning-moses/#comment-87 chhean Mon, 20 Feb 2012 22:12:00 +0000 http://blog.opencog.org/?p=326#comment-87 Well done Linas Well done Linas

]]> Comment on Determining word senses from grammatical usage by Rob http://blog.opencog.org/2009/01/12/determining-word-senses-from-grammatical-usage/#comment-86 Rob Tue, 14 Feb 2012 10:19:00 +0000 http://brainwave.opencog.org/?p=70#comment-86 Hi, So how have you progressed with this model? Hi,

So how have you progressed with this model?

]]> Comment on Tuning Metalearning in MOSES by J P http://blog.opencog.org/2012/02/07/tuning-moses/#comment-85 J P Mon, 13 Feb 2012 00:26:00 +0000 http://blog.opencog.org/?p=326#comment-85 Very nice write up Linas - thanks :-) Very nice write up Linas – thanks :-)

]]> Comment on Preview of a virtual learning environment by MDude1350 http://blog.opencog.org/2011/08/04/virtual-learning-environment/#comment-83 MDude1350 Tue, 10 Jan 2012 05:11:00 +0000 http://blog.opencog.org/?p=313#comment-83 That Bomberman Hero music. This virtual learning environment sounds like something I've been wanting for a while. The bots you have seem nice, but I think it'd be fun to see what I can make on my own. Would I be able to use the environment to embody other AI systems without them being  a part of OpenCog, or would I need to roll my own system for that? That Bomberman Hero music.
This virtual learning environment sounds like something I’ve been wanting for a while. The bots you have seem nice, but I think it’d be fun to see what I can make on my own. Would I be able to use the environment to embody other AI systems without them being  a part of OpenCog, or would I need to roll my own system for that?

]]> Comment on Preview of a virtual learning environment by J P http://blog.opencog.org/2011/08/04/virtual-learning-environment/#comment-78 J P Sat, 17 Sep 2011 07:55:00 +0000 http://blog.opencog.org/?p=313#comment-78 There is definitely an aspect like curiousity. In OpenPsi (the emotion and motivation system), there are demands for "competence" and "certainty". "Certainty" will lead the character to try to understand the state of environment, "competence" will lead to the character trying to get better at influence and predict the result of its actions. There is definitely an aspect like curiousity. In OpenPsi (the emotion and motivation system), there are demands for “competence” and “certainty”. “Certainty” will lead the character to try to understand the state of environment, “competence” will lead to the character trying to get better at influence and predict the result of its actions.

]]> Comment on Preview of a virtual learning environment by Sa3vis http://blog.opencog.org/2011/08/04/virtual-learning-environment/#comment-77 Sa3vis Fri, 16 Sep 2011 14:15:00 +0000 http://blog.opencog.org/?p=313#comment-77 I'm just wondering if curiosity and memory will be added in this project. Curiosity in that if there is a new object or button it will go observe it and then depending on what that does it will either ignore, stay away from, be cautious with or use. Maybe even hoard. And with memory then if it sees something it doesn't need at the moment it  would keep a virtual memory map of where it is located and then find it when needed. Those are just a few things I was thinking would be neat. Keep up the great work and good luck. I’m just wondering if curiosity and memory will be added in this project. Curiosity in that if there is a new object or button it will go observe it and then depending on what that does it will either ignore, stay away from, be cautious with or use. Maybe even hoard. And with memory then if it sees something it doesn’t need at the moment it  would keep a virtual memory map of where it is located and then find it when needed. Those are just a few things I was thinking would be neat. Keep up the great work and good luck.

]]> Comment on OpenCog Hong Kong Project by Matt http://blog.opencog.org/2010/12/07/opencog-hong-kong-project/#comment-74 Matt Wed, 24 Aug 2011 20:12:00 +0000 http://blog.opencog.org/?p=253#comment-74 Wow, I have used many of the tools that you have (festival, zeromq, [amqp], [rabbitmq], [xmpp], protocol buffers, [raw sockets in raw php or etc], .. etc;;;. Wow, I have used many of the tools that you have (festival, zeromq, [amqp], [rabbitmq], [xmpp], protocol buffers, [raw sockets in raw php or etc], .. etc;;;.

]]> Comment on Preview of a virtual learning environment by J P http://blog.opencog.org/2011/08/04/virtual-learning-environment/#comment-73 J P Thu, 11 Aug 2011 04:45:00 +0000 http://blog.opencog.org/?p=313#comment-73 Hi - My aim is to be as transparent as possible about what we're doing. We're not interested in trying to deceive people with narrow scripted behaviours. So, saying that, these videos are mostly a demonstration of the virtual world. OpenCog has specialised path-finding algorithms, but these can be constructed as part of a "action plan". So a plan might include an action to "move to the battery cube" (where battery cube is a unique id related to a previously seen object). So we're abstracting actions at a certain level, but not confining OpenCog to only know about certain objects. We can easily add new actions to objects and it's up to OpenCog to detect the patterns of what happens when those actions are initiated. We are building on an older embodiment system that was quite limited in what it could learn. Now a lot of time is being spent on pulling out the old parts and replacing them with general systems. Hi – My aim is to be as transparent as possible about what we’re doing. We’re not interested in trying to deceive people with narrow scripted behaviours.

So, saying that, these videos are mostly a demonstration of the virtual world. OpenCog has specialised path-finding algorithms, but these can be constructed as part of a “action plan”. So a plan might include an action to “move to the battery cube” (where battery cube is a unique id related to a previously seen object).

So we’re abstracting actions at a certain level, but not confining OpenCog to only know about certain objects. We can easily add new actions to objects and it’s up to OpenCog to detect the patterns of what happens when those actions are initiated.

We are building on an older embodiment system that was quite limited in what it could learn. Now a lot of time is being spent on pulling out the old parts and replacing them with general systems.

]]>