Loading ...
Sorry, an error occurred while loading the content.

Outer Space Robots WOOT!

Expand Messages
  • connor_ramsey@ymail.com
    You know, robotics isn t my only nag. I also have quite a thing for antigravity devices. Why? Because they re COOL. But there s actually a pretty simple way to
    Message 1 of 23 , Oct 16, 2012
    • 0 Attachment

      You know, robotics isn't my only nag. I also have quite a thing for antigravity devices. Why? Because they're COOL. But there's actually a pretty simple way to achieve antigravity: have two or three donut shaped permanent magnets( three along the rim is probably more stable than two in the center), and have three DC motors. have the motors positioned along the rim angled in or out, and the magnets fixed to the axles. I think the reason you need to use DC current might be because the torrent field(or something) would interfere with AC, I dunno. Also, a three motor antigravity device could allow a BEAM circuit to individually manage each motor. As far as anything, the circuitry simply needs to dynamically stabilize the device, and afterward some sort of behavior can be build around that in the form of deliberate imbalancing to create inertial thrust.

      I know I'm getting in over everyone's heads here, but it's actually pretty simple stuff, mechanically anyways. I'm sure that a levitating BEAMbot is fairly within the bounderies of BEAM. The circuitry would be simple. The mechanics would be simple.

      Actually I think OISM might be more well suited to balancing/stabilizing applications than traditional BEAM. So perhaps I'll keep this in store for all to see eventually.

      Hope you enjoy, Connor

    • Amit
      Sounds like a cooler version of a magbot. I love desk toys.
      Message 2 of 23 , Oct 16, 2012
      • 0 Attachment
        Sounds like a cooler version of a magbot. I love desk toys.

        --- In beam@yahoogroups.com, "connor_ramsey@..." <connor_ramsey@...> wrote:
        >
        >
        > You know, robotics isn't my only nag. I also have quite a thing for
        > antigravity devices. Why? Because they're COOL. But there's actually a
        > pretty simple way to achieve antigravity: have two or three donut shaped
        > permanent magnets( three along the rim is probably more stable than two
        > in the center), and have three DC motors. have the motors positioned
        > along the rim angled in or out, and the magnets fixed to the axles. I
        > think the reason you need to use DC current might be because the torrent
        > field(or something) would interfere with AC, I dunno. Also, a three
        > motor antigravity device could allow a BEAM circuit to individually
        > manage each motor. As far as anything, the circuitry simply needs to
        > dynamically stabilize the device, and afterward some sort of behavior
        > can be build around that in the form of deliberate imbalancing to create
        > inertial thrust.
        >
        > I know I'm getting in over everyone's heads here, but it's actually
        > pretty simple stuff, mechanically anyways. I'm sure that a levitating
        > BEAMbot is fairly within the bounderies of BEAM. The circuitry would be
        > simple. The mechanics would be simple.
        >
        > Actually I think OISM might be more well suited to balancing/stabilizing
        > applications than traditional BEAM. So perhaps I'll keep this in store
        > for all to see eventually.
        >
        > Hope you enjoy, Connor
        >
      • David Buckley
        Connor Do you have a reference link for the doughnut magnet antigravity device or is it your own idea? Has one ever worked? OISM is what? David ... From:
        Message 3 of 23 , Oct 16, 2012
        • 0 Attachment
          Connor
          Do you have a reference link for the doughnut magnet antigravity device or is it your own idea? Has one ever worked?
          OISM is what?
          David
           
          ----- Original Message -----
          Sent: Tuesday, October 16, 2012 10:02 AM
          Subject: [beam] Outer Space Robots WOOT!

           

          You know, robotics isn't my only nag. I also have quite a thing for antigravity devices. Why? Because they're COOL. But there's actually a pretty simple way to achieve antigravity: have two or three donut shaped permanent magnets( three along the rim is probably more stable than two in the center), and have three DC motors. have the motors positioned along the rim angled in or out, and the magnets fixed to the axles. I think the reason you need to use DC current might be because the torrent field(or something) would interfere with AC, I dunno. Also, a three motor antigravity device could allow a BEAM circuit to individually manage each motor. As far as anything, the circuitry simply needs to dynamically stabilize the device, and afterward some sort of behavior can be build around that in the form of deliberate imbalancing to create inertial thrust.

          I know I'm getting in over everyone's heads here, but it's actually pretty simple stuff, mechanically anyways. I'm sure that a levitating BEAMbot is fairly within the bounderies of BEAM. The circuitry would be simple. The mechanics would be simple.

          Actually I think OISM might be more well suited to balancing/stabilizing applications than traditional BEAM. So perhaps I'll keep this in store for all to see eventually.

          Hope you enjoy, Connor

        • connor_ramsey@ymail.com
          ... Type in gyroscopic antigravity on youtube or google it, and its one of the first things to pop up. There s more than one video, and the device shown
          Message 4 of 23 , Oct 16, 2012
          • 0 Attachment
            --- In beam@yahoogroups.com, "David Buckley" <david@...> wrote:
            >
            > Connor
            > Do you have a reference link for the doughnut magnet antigravity device or is it your own idea? Has one ever worked?
            > OISM is what?
            > David


            Type in gyroscopic antigravity on youtube or google it, and its one of the first things to pop up. There's more than one video, and the device shown obviously has a lot of room for improvement. Really, its just a proof of concept. My version is a list of improvements really.

            And yeah, the idea's basically a levitating mobile magbot.

            ANd if you don't know what OISM is, there's a nice, comprehensive page for it on the wiki.
          • Amit
            If it the video with the lovely contraption consisting of those off axis gyros rigidly mounted to a principal axis about a cage, then that s not antigravity.
            Message 5 of 23 , Oct 17, 2012
            • 0 Attachment
              If it' the video with the lovely contraption consisting of those off axis gyros rigidly mounted to a principal axis about a cage, then that's not antigravity. That's just conservation of angular momentum and inertial coupling effects.

              I thought it would be something with an object floating on a magnetic field generated by three slightly off axis coils... Oh well



              --- In beam@yahoogroups.com, "connor_ramsey@..." <connor_ramsey@...> wrote:
              >
              >
              >
              > --- In beam@yahoogroups.com, "David Buckley" <david@> wrote:
              > >
              > > Connor
              > > Do you have a reference link for the doughnut magnet antigravity device or is it your own idea? Has one ever worked?
              > > OISM is what?
              > > David
              >
              >
              > Type in gyroscopic antigravity on youtube or google it, and its one of the first things to pop up. There's more than one video, and the device shown obviously has a lot of room for improvement. Really, its just a proof of concept. My version is a list of improvements really.
              >
              > And yeah, the idea's basically a levitating mobile magbot.
              >
              > ANd if you don't know what OISM is, there's a nice, comprehensive page for it on the wiki.
              >
            • David Buckley
              OISM that, I don t think humans have big enough brains to create OISM circuits to do anything useful.Chaotic systems are really fascinating but too complex to
              Message 6 of 23 , Oct 17, 2012
              • 0 Attachment
                OISM that, I don't think humans have big enough brains to create OISM circuits to do anything useful.Chaotic systems are really fascinating but too complex to comprehend unless of course you are someone like Srinvasa Ramanujan.
                I saw the Laithwaite's original TV lectures where he demonstrated Gyros however the effects don't hppen unless one end of the shaft is restrained and the wheel can move in a gravitational field.
                What has it got to do with magnets?
                The Gyro demonstration requires energy which is extracted from the inertial energy of the wheel. It is no free lunch.
                David
                 
                ----- Original Message -----
                Sent: Wednesday, October 17, 2012 5:34 AM
                Subject: [beam] Re: Outer Space Robots WOOT!

                 



                --- In beam@yahoogroups.com, "David Buckley" <david@...> wrote:
                >
                > Connor
                > Do you have a reference link for the doughnut magnet antigravity device or is it your own idea? Has one ever worked?
                > OISM is what?
                > David

                Type in gyroscopic antigravity on youtube or google it, and its one of the first things to pop up. There's more than one video, and the device shown obviously has a lot of room for improvement. Really, its just a proof of concept. My version is a list of improvements really.

                And yeah, the idea's basically a levitating mobile magbot.

                ANd if you don't know what OISM is, there's a nice, comprehensive page for it on the wiki.

              • connor_ramsey@ymail.com
                ... Yeah that one. It s just proofing, but there definitely seems to be extensive weight loss due to the effect. If you look at the second video, you can see
                Message 7 of 23 , Oct 17, 2012
                • 0 Attachment
                  --- In beam@yahoogroups.com, "Amit" <amitjones101@...> wrote:
                  >
                  > If it' the video with the lovely contraption consisting of those off axis gyros rigidly mounted to a principal axis about a cage, then that's not antigravity. That's just conservation of angular momentum and inertial coupling effects.
                  >
                  > I thought it would be something with an object floating on a magnetic field generated by three slightly off axis coils... Oh well

                  Yeah that one. It's just proofing, but there definitely seems to be extensive weight loss due to the effect. If you look at the second video, you can see the improved device is capable of levitation, in a way, gravity neutral. It just floats. Note that there does seem to be at least one wire on the rim of the vehichle to stabilize it, but also note there's no visible tension of the wire. If it was indeed supporting the device's weight, it would certainly not be so loose. And the device itself would be toppled on its side, considering the wire is only on the rim.

                  So there's definitely some sort of inertial propulsion going on there, whether it's mechanical or magnetic, or likely both.

                  But back to the world of robotics, I was thinking of a robot that would survive just fine in the wild for an extended period of time without any human interaction whatsoever. That's kind of the ideal model in BEAM. But unfortunately the robot would eventually corode, decay, and die. But what if the robot was equipped with a personal ionizer, that negatively ionized the air around the robot, which would completely pacify all chemical reactions between the bot's parts and the atmosphere. How long would it last then?

                  They did the same to the Jesus statue over Rio de Janeiro back in 2002. As long as the power stays on, the statue will NEVER corrode. Ever. Because since chemical interations are all about giving and taking electrons, and since negative ions with full outer shells can do neither, it might as well be surrounded with a noble gas.

                  So how long might a robot last this way. Maybe indefinitely. Maybe not. Just because it never corrodes doesn't mean it won't stumble into danger. But as long as the machine is adept enough to effectively deal with such circumstances, and as long as it remains operational, it should remain just fine for a very, very long time.
                  Maybe on the order of centuries, millenia even. Who knows.

                  Hope you enjoy,
                  Connor
                • David Buckley
                  Connor If it is the device which falls over at the end in video part 2 he doesn t understand the effects of vibration, friction in the motor bearings or most
                  Message 8 of 23 , Oct 18, 2012
                  • 0 Attachment
                    Connor
                    If it is the device which falls over at the end in video part 2 he doesn't understand the effects of vibration, friction in the motor bearings or most anything about making stuff. 4 out of 10 for effort, 1 out of 10 for ability, he needs to keep up the art and design classes ;-).
                    Of course that may not be the video you referred to but without reference links it is hard to tell.
                    (no link here because I don't want Google finding links to those videos.)
                     
                    There are more important points to consider than decay after a few hundred/thousand years. One would be motor life. If it has brushes you are looking at about a few thousand hours brush life maximum. If it has gears you are looking at about the same lifetime. even if it doesn't move a lot the oils in bearings will dry up etc.
                     
                    Let's be realistic, if BEAM itself is to survive. let alone the robots. it has to show BEAM robots can do more than stumble a few steps or track the sun or oscillate a bit etc.
                    So far even including Mark's enthusiastic showing of his walkers we don't have much evidence of that. The circuits have elegance but BEAM robots don't do much. They are not now not competitive with simple program controlled robots - most programs of which are certainly not elegant!
                    Just as a traditional Beam controller is a physical (electronic gates etc) structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions, so a brain can be built in software which is a virtual structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions.
                    The principles are the same but the implementation is different. How many neurons could you build in an 8 pin PIC, maybe 5 or 10, lets stick with 5, Then with ten chips (80 pins) you could have 50 neurons. Now how about your cell phone, how many neurons could you build in its processors?
                    An ordinary house fly has about 1 million neurons. Neurons count when it comes to ability. And even house flies don't survive very long, not even close to the 600 years of Richard Daniel (Clifford D.Simak. All the traps of Earth).
                    There is nothing horrific about building brains in software. Take biology, start with some atoms and arrange them into molecules, then arrange lots of different types of molecules into cells, then arrange quite a few different types of cells into brains, then inject the electric currents/fields which keep the organism alive, then inject more information - modulations of those electric currents/fields which enable the organism to do something. It is all a hierarchy
                    At one time building BEAM robots was a good introduction to robotics but that is no longer true, BEAM is on a path to nowhere unless a way is found of scaling up the number and types of neurons  and functionality of neuron clusters.
                    When that is accomplished then maybe work could begin on emulating the Pharaohs with a life of millions of years.
                    David
                     
                     
                     
                     
                     
                     
                     
                     
                     
                     
                     
                     
                     
                    ----- Original Message -----
                    Sent: Thursday, October 18, 2012 7:58 AM
                    Subject: [beam] Re: Outer Space Robots WOOT!

                     



                    --- In beam@yahoogroups.com, "Amit" <amitjones101@...> wrote:
                    >
                    > If it' the video with the lovely contraption consisting of those off axis gyros rigidly mounted to a principal axis about a cage, then that's not antigravity. That's just conservation of angular momentum and inertial coupling effects.
                    >
                    > I thought it would be something with an object floating on a magnetic field generated by three slightly off axis coils... Oh well

                    Yeah that one. It's just proofing, but there definitely seems to be extensive weight loss due to the effect. If you look at the second video, you can see the improved device is capable of levitation, in a way, gravity neutral. It just floats. Note that there does seem to be at least one wire on the rim of the vehichle to stabilize it, but also note there's no visible tension of the wire. If it was indeed supporting the device's weight, it would certainly not be so loose. And the device itself would be toppled on its side, considering the wire is only on the rim.

                    So there's definitely some sort of inertial propulsion going on there, whether it's mechanical or magnetic, or likely both.

                    But back to the world of robotics, I was thinking of a robot that would survive just fine in the wild for an extended period of time without any human interaction whatsoever. That's kind of the ideal model in BEAM. But unfortunately the robot would eventually corode, decay, and die. But what if the robot was equipped with a personal ionizer, that negatively ionized the air around the robot, which would completely pacify all chemical reactions between the bot's parts and the atmosphere. How long would it last then?

                    They did the same to the Jesus statue over Rio de Janeiro back in 2002. As long as the power stays on, the statue will NEVER corrode. Ever. Because since chemical interations are all about giving and taking electrons, and since negative ions with full outer shells can do neither, it might as well be surrounded with a noble gas.

                    So how long might a robot last this way. Maybe indefinitely. Maybe not. Just because it never corrodes doesn't mean it won't stumble into danger. But as long as the machine is adept enough to effectively deal with such circumstances, and as long as it remains operational, it should remain just fine for a very, very long time.
                    Maybe on the order of centuries, millenia even. Who knows.

                    Hope you enjoy,
                    Connor

                  • Richard Piotter
                    I m gonna half and half this on my next robot. I m still gonna use analog Nv circuits, but they will be paired to xilinx CPLD chips, so the connectome, or
                    Message 9 of 23 , Oct 18, 2012
                    • 0 Attachment
                      I'm gonna half and half this on my next robot. I'm still gonna use
                      analog Nv circuits, but they will be paired to xilinx CPLD chips, so
                      the connectome, or network map can be easily modified by PC, and even
                      made dynamic, changing when certain inputs are received.

                      Saves me the trouble of having to figure out how to simulate the
                      neuron in code, but still gives me the dynamic structure affordable in
                      software!

                      Richard Piotter


                      Begin forwarded message:
                      > From: "David Buckley" <david@...>
                      > Date: October 18, 2012 7:36:02 AM CDT
                      > To: <beam@yahoogroups.com>
                      > Subject: Re: [beam] Re: Outer Space Robots WOOT!
                      > Reply-To: beam@yahoogroups.com
                      >
                      >
                      > Connor
                      > If it is the device which falls over at the end in video part 2 he
                      > doesn't understand the effects of vibration, friction in the motor
                      > bearings or most anything about making stuff. 4 out of 10 for
                      > effort, 1 out of 10 for ability, he needs to keep up the art and
                      > design classes ;-).
                      > Of course that may not be the video you referred to but without
                      > reference links it is hard to tell.
                      > (no link here because I don't want Google finding links to those
                      > videos.)
                      >
                      > There are more important points to consider than decay after a few
                      > hundred/thousand years. One would be motor life. If it has brushes
                      > you are looking at about a few thousand hours brush life maximum. If
                      > it has gears you are looking at about the same lifetime. even if it
                      > doesn't move a lot the oils in bearings will dry up etc.
                      >
                      > Let's be realistic, if BEAM itself is to survive. let alone the
                      > robots. it has to show BEAM robots can do more than stumble a few
                      > steps or track the sun or oscillate a bit etc.
                      > So far even including Mark's enthusiastic showing of his walkers we
                      > don't have much evidence of that. The circuits have elegance but
                      > BEAM robots don't do much. They are not now not competitive with
                      > simple program controlled robots - most programs of which are
                      > certainly not elegant!
                      > Just as a traditional Beam controller is a physical (electronic
                      > gates etc) structure in which information (what is in the mind of
                      > the robot) ebbs and flows and controls the robot's actions, so a
                      > brain can be built in software which is a virtual structure in which
                      > information (what is in the mind of the robot) ebbs and flows and
                      > controls the robot's actions.
                      > The principles are the same but the implementation is different. How
                      > many neurons could you build in an 8 pin PIC, maybe 5 or 10, lets
                      > stick with 5, Then with ten chips (80 pins) you could have 50
                      > neurons. Now how about your cell phone, how many neurons could you
                      > build in its processors?
                      > An ordinary house fly has about 1 million neurons. Neurons count
                      > when it comes to ability. And even house flies don't survive very
                      > long, not even close to the 600 years of Richard Daniel (Clifford
                      > D.Simak. All the traps of Earth).
                      > There is nothing horrific about building brains in software. Take
                      > biology, start with some atoms and arrange them into molecules, then
                      > arrange lots of different types of molecules into cells, then
                      > arrange quite a few different types of cells into brains, then
                      > inject the electric currents/fields which keep the organism alive,
                      > then inject more information - modulations of those electric
                      > currents/fields which enable the organism to do something. It is all
                      > a hierarchy
                      > At one time building BEAM robots was a good introduction to robotics
                      > but that is no longer true, BEAM is on a path to nowhere unless a
                      > way is found of scaling up the number and types of neurons and
                      > functionality of neuron clusters.
                      > When that is accomplished then maybe work could begin on emulating
                      > the Pharaohs with a life of millions of years.
                      > David
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      >
                      > ----- Original Message -----
                      > From: connor_ramsey@...
                      > To: beam@yahoogroups.com
                      > Sent: Thursday, October 18, 2012 7:58 AM
                      > Subject: [beam] Re: Outer Space Robots WOOT!
                      >
                      >
                      >
                      >
                      > --- In beam@yahoogroups.com, "Amit" <amitjones101@...> wrote:
                      > >
                      > > If it' the video with the lovely contraption consisting of those
                      > off axis gyros rigidly mounted to a principal axis about a cage,
                      > then that's not antigravity. That's just conservation of angular
                      > momentum and inertial coupling effects.
                      > >
                      > > I thought it would be something with an object floating on a
                      > magnetic field generated by three slightly off axis coils... Oh well
                      >
                      > Yeah that one. It's just proofing, but there definitely seems to be
                      > extensive weight loss due to the effect. If you look at the second
                      > video, you can see the improved device is capable of levitation, in
                      > a way, gravity neutral. It just floats. Note that there does seem to
                      > be at least one wire on the rim of the vehichle to stabilize it,
                      > but also note there's no visible tension of the wire. If it was
                      > indeed supporting the device's weight, it would certainly not be so
                      > loose. And the device itself would be toppled on its side,
                      > considering the wire is only on the rim.
                      >
                      > So there's definitely some sort of inertial propulsion going on
                      > there, whether it's mechanical or magnetic, or likely both.
                      >
                      > But back to the world of robotics, I was thinking of a robot that
                      > would survive just fine in the wild for an extended period of time
                      > without any human interaction whatsoever. That's kind of the ideal
                      > model in BEAM. But unfortunately the robot would eventually corode,
                      > decay, and die. But what if the robot was equipped with a personal
                      > ionizer, that negatively ionized the air around the robot, which
                      > would completely pacify all chemical reactions between the bot's
                      > parts and the atmosphere. How long would it last then?
                      >
                      > They did the same to the Jesus statue over Rio de Janeiro back in
                      > 2002. As long as the power stays on, the statue will NEVER corrode.
                      > Ever. Because since chemical interations are all about giving and
                      > taking electrons, and since negative ions with full outer shells can
                      > do neither, it might as well be surrounded with a noble gas.
                      >
                      > So how long might a robot last this way. Maybe indefinitely. Maybe
                      > not. Just because it never corrodes doesn't mean it won't stumble
                      > into danger. But as long as the machine is adept enough to
                      > effectively deal with such circumstances, and as long as it remains
                      > operational, it should remain just fine for a very, very long time.
                      > Maybe on the order of centuries, millenia even. Who knows.
                      >
                      > Hope you enjoy,
                      > Connor
                      >
                      >
                      >
                      > <!-- #ygrp-mkp { border: 1px solid #d8d8d8; font-family: Arial;
                      > margin: 10px 0; padding: 0 10px; } #ygrp-mkp hr { border: 1px solid
                      > #d8d8d8; } #ygrp-mkp #hd { color: #628c2a; font-size: 85%; font-
                      > weight: 700; line-height: 122%; margin: 10px 0; } #ygrp-mkp #ads
                      > { margin-bottom: 10px; } #ygrp-mkp .ad { padding: 0 0; } #ygrp-
                      > mkp .ad p { margin: 0; } #ygrp-mkp .ad a { color: #0000ff; text-
                      > decoration: none; } #ygrp-sponsor #ygrp-lc { font-family: Arial; }
                      > #ygrp-sponsor #ygrp-lc #hd { margin: 10px 0px; font-weight: 700;
                      > font-size: 78%; line-height: 122%; } #ygrp-sponsor #ygrp-lc .ad
                      > { margin-bottom: 10px; padding: 0 0; } a { color: #1e66ae; }
                      > #actions { font-family: Verdana; font-size: 11px; padding: 10px 0; }
                      > #activity { background-color: #e0ecee; float: left; font-family:
                      > Verdana; font-size: 10px; padding: 10px; } #activity span { font-
                      > weight: 700; } #activity span:first-child { text-transform:
                      > uppercase; } #activity span a { color: #5085b6; text-decoration:
                      > none; } #activity span span { color: #ff7900; } #activity
                      > span .underline { text-decoration: underline; } .attach { clear:
                      > both; display: table; font-family: Arial; font-size: 12px; padding:
                      > 10px 0; width: 400px; } .attach div a { text-decoration:
                      > none; } .attach img { border: none; padding-right: 5px; } .attach
                      > label { display: block; margin-bottom: 5px; } .attach label a { text-
                      > decoration: none; } blockquote { margin: 0 0 0 4px; } .bold { font-
                      > family: Arial; font-size: 13px; font-weight: 700; } .bold a { text-
                      > decoration: none; } dd.last p a { font-family: Verdana; font-weight:
                      > 700; } dd.last p span { margin-right: 10px; font-family: Verdana;
                      > font-weight: 700; } dd.last p span.yshortcuts { margin-right:
                      > 0; } div.attach-table div div a { text-decoration: none; }
                      > div.attach-table { width: 400px; } div.file-title a, div.file-title
                      > a:active, div.file-title a:hover, div.file-title a:visited { text-
                      > decoration: none; } div.photo-title a, div.photo-title a:active,
                      > div.photo-title a:hover, div.photo-title a:visited { text-
                      > decoration: none; } div#ygrp-mlmsg #ygrp-msg p a span.yshortcuts
                      > { font-family: Verdana; font-size: 10px; font-weight:
                      > normal; } .green { color: #628c2a; } .MsoNormal { margin: 0 0 0 0; }
                      > o { font-size: 0; } #photos div { float: left; width: 72px; }
                      > #photos div div { border: 1px solid #666666; height: 62px; overflow:
                      > hidden; width: 62px; } #photos div label { color: #666666; font-
                      > size: 10px; overflow: hidden; text-align: center; white-space:
                      > nowrap; width: 64px; } #reco-category { font-size: 77%; } #reco-
                      > desc { font-size: 77%; } .replbq { margin: 4px; } #ygrp-actbar div
                      > a:first-child { /* border-right: 0px solid #000;*/ margin-right:
                      > 2px; padding-right: 5px; } #ygrp-mlmsg { font-size: 13px; font-
                      > family: Arial, helvetica,clean, sans-serif; *font-size: small;
                      > *font: x-small; } #ygrp-mlmsg table { font-size: inherit; font:
                      > 100%; } #ygrp-mlmsg select, input, textarea { font: 99% Arial,
                      > Helvetica, clean, sans-serif; } #ygrp-mlmsg pre, code { font:115%
                      > monospace; *font-size:100%; } #ygrp-mlmsg * { line-height: 1.22em; }
                      > #ygrp-mlmsg #logo { padding-bottom: 10px; } #ygrp-mlmsg a { color:
                      > #1E66AE; } #ygrp-msg p a { font-family: Verdana; } #ygrp-msg
                      > p#attach-count span { color: #1E66AE; font-weight: 700; } #ygrp-reco
                      > #reco-head { color: #ff7900; font-weight: 700; } #ygrp-reco { margin-
                      > bottom: 20px; padding: 0px; } #ygrp-sponsor #ov li a { font-size:
                      > 130%; text-decoration: none; } #ygrp-sponsor #ov li { font-size:
                      > 77%; list-style-type: square; padding: 6px 0; } #ygrp-sponsor #ov ul
                      > { margin: 0; padding: 0 0 0 8px; } #ygrp-text { font-family:
                      > Georgia; } #ygrp-text p { margin: 0 0 1em 0; } #ygrp-text tt { font-
                      > size: 120%; } #ygrp-vital ul li:last-child { border-right: none !
                      > important; } -->
                    • connor_ramsey@ymail.com
                      ... doesn t understand the effects of vibration, friction in the motor bearings or most anything about making stuff. 4 out of 10 for effort, 1 out of 10 for
                      Message 10 of 23 , Oct 18, 2012
                      • 0 Attachment


                        --- In beam@yahoogroups.com, "David Buckley" <david@...> wrote:
                        >
                        > Connor
                        > If it is the device which falls over at the end in video part 2 he doesn't understand the effects of vibration, friction in the motor bearings or most anything about making stuff. 4 out of 10 for effort, 1 out of 10 for ability, he needs to keep up the art and design classes ;-).
                        > Of course that may not be the video you referred to but without reference links it is hard to tell.
                        > (no link here because I don't want Google finding links to those videos.)
                        >
                        > There are more important points to consider than decay after a few hundred/thousand years. One would be motor life. If it has brushes you are looking at about a few thousand hours brush life maximum. If it has gears you are looking at about the same lifetime. even if it doesn't move a lot the oils in bearings will dry up etc.
                        >
                        > Let's be realistic, if BEAM itself is to survive. let alone the robots. it has to show BEAM robots can do more than stumble a few steps or track the sun or oscillate a bit etc.
                        > So far even including Mark's enthusiastic showing of his walkers we don't have much evidence of that. The circuits have elegance but BEAM robots don't do much. They are not now not competitive with simple program controlled robots - most programs of which are certainly not elegant!
                        > Just as a traditional Beam controller is a physical (electronic gates etc) structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions, so a brain can be built in software which is a virtual structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions.
                        > The principles are the same but the implementation is different. How many neurons could you build in an 8 pin PIC, maybe 5 or 10, lets stick with 5, Then with ten chips (80 pins) you could have 50 neurons. Now how about your cell phone, how many neurons could you build in its processors?
                        > An ordinary house fly has about 1 million neurons. Neurons count when it comes to ability. And even house flies don't survive very long, not even close to the 600 years of Richard Daniel (Clifford D.Simak. All the traps of Earth).
                        > There is nothing horrific about building brains in software. Take biology, start with some atoms and arrange them into molecules, then arrange lots of different types of molecules into cells, then arrange quite a few different types of cells into brains, then inject the electric currents/fields which keep the organism alive, then inject more information - modulations of those electric currents/fields which enable the organism to do something. It is all a hierarchy
                        > At one time building BEAM robots was a good introduction to robotics but that is no longer true, BEAM is on a path to nowhere unless a way is found of scaling up the number and types of neurons and functionality of neuron clusters.
                        > When that is accomplished then maybe work could begin on emulating the Pharaohs with a life of millions of years.
                        > David
                        >

                        YES! Yes yes yes yes yes yes yes! That's exactly the case with BEAM. The problem in my mind is the KISS rule. By that logic, KISS itself, by which almost every consecutive BEAMer follows, is the very path to BEAM's ultimate destruction.

                        However, I don't believe the solution is necessarily in software. Why? Because while superficially software seems to have a very large capacitative advantage over traditional BEAM, which it does, everyone simply fails to make a relative comparison between the two technologies. For example: A typical microcontroller used in small robots, such as that Atmel chip on Solarbotics, contains several million transistors. Now one can virtually construct up to perhaps a 10 neuron system on the chip, having many advantages over hardware built circuits such as convenience, flexibility, reconfigureability, etc. A PC has on average several billion transistors, some of the newer ones up to several trillion transistors. Now how many virtual neurons can that fit onto it? Thousands if not tens or hundreds of thousands of complex interneural pathways and action potentials would be buzzing through that software.

                        Now take physical hardware based neural networks. Traditional BEAM only goes so far as simple coupled oscillators on average, rarely something beyond that, and quite often something borderline to a Braitenburg vehichle. However, what if there was a chip of identical size to a microcontroller and the same number of transistors, but is constructed using a neural architecture rather than a digital one, and each inverter represents a single neuron out of millions of transistors.

                        So I'll let the question sink in before I go: Does the digital architecture really exceed BEAM performance-wise when BEAM is scaled up, or down really, to the same level of complexity? The same goes for scaling down the digital architecture to BEAM's level. In either case. Which system is more capable?

                         

                        P.S. yes that is the video. It may work at a substandard level, but it's only a proof of concept. It's not really supposed to work well quite yet. But it definitely does more than nothing. That's what's important.

                        Connor

                      • David Buckley
                        Connor A well yes if only, this existed or that existed, but the sort of chips you are talking about don t. I guess you could fashion them from Gate Arrays
                        Message 11 of 23 , Oct 18, 2012
                        • 0 Attachment
                          Connor
                          A well yes if only, this existed or that existed, but the sort of chips you are talking about don't. I guess you could fashion them from Gate Arrays however there is still the problem of just how do you wire them all up, even virtually? What sort of schema are you going to follow?
                          Forget the number of transistors on a chip, there are actually quite a lot in the 74 and 4000 series chips, but who cares.
                          What the chips do is what is important, wetware brains have quite a lot of cells that aren't neurons, actually more in fact, and each cell has a heck of a lot of constituent parts, so counting transistors is pointless, you might as well count proteins in a brain.
                          The thing about software is you can play around with much more complicated systems without any soldering.
                          Having figured out what you can do then it could be implemented on a Gate Array (I guess, as I have never used any). There is also the Green Arrays Forth chip. However they have development overheads, like money and time to learn.
                          I don't know what chips SolarBotics sell and or which you meant.
                          This last year by accident I built another biped walker Bambino
                           
                          It can walk forwards and backwards and turn right or left;
                          it walks by lifting its feet off the floor and doesn't use overlapping feet as toy robots do;
                          it is driven by two servos;
                          it is powered by 3 AA cells in a holder (can be changed to 4 AA cells to suit 5v electronics such as the Arduino);
                          it is teachable from an Infra-Red handset (Sony TV remote);
                          it can be programmed in Basic; and uses a PicAxe 20M2
                          the high level program has built in routines to :
                          - listen out for commands from the handset;
                          - listen out for commands through the programming/serial port (see below)
                          - avoid obstacles
                          - avoid drop-offs, ie table edges;
                          - walk to a light;
                          - walk away from a light;
                          - (projected are to be able to follow a wall for a distance etc)
                          it has :
                          - two forward facing optical obstacle sensors;
                          - two downward facing optical obstacle sensors to sense drop-offs, ie table edges;
                          - a piezo sounder for buzzes and tunes;
                          - an indicator LED;
                          - two high power LEDs which can be PWM controlled and used for HMI (Human Machine Interface)
                          - two other high power LEDs which can be used for HMI;
                          - an IR receiver so it can respond to signals from a IR remote handset, or other robots;
                          - an IR-LED so it can talk to other robots using the same commands as from the handset, or using extended codes;
                          - a connector for an Ultrasonic ranging module;
                          - a connector for a compass module;
                          - A serial programming connector which can also be used to talk to an add-on higher level processor such as an Arduino;
                          it can;
                          - learn and remember up to 8 (will be 16) sequences from the handset;
                          -- each sequence can chain to another sequence or itself so loops can be created;
                          -- any sequence can call any Basic subroutine
                          -- Basic subroutines will be able to invoke any learned sequence;
                          it has : - a complete suit of programs taking users from flashing an LED to making the robot walk while avoiding obstacle etc, now up to #28
                           
                          The chip is about £2, say $3 or $4 and that is all you need, no crystals or anything.
                          The software isn't BEAM inspired but it could be although it would loose a lot of functionality because of inbuilt interfaces for IRcodes etc.
                           
                          Now it is hard enough to pack all that in without delving into chaos theory as well
                          We know that creating large versatile neural nets is very difficult because nobody has managed to do it yet.
                          Unless there is some plan for BEAM brains which can be shown to work NOBODY is going to be interested in creating chips.
                          The software for Bambino is conventional but the software running my other bipeds isn't. That software builds the structure of a brain which then process robot command messages - codes - and it is the codes which are the program if you like, much like the impulses circulating in nervous nets.
                          At the end of the day I want functionality not adherence to some dogma.
                           
                           
                          Personally I don't think the video shows anything and neither do any of the other videos on YouTube except that they show that non of the makers know what they are talking about, even those who are trying to explain gyros from conventional physics haven't a clue even about conventional physics let alone gyros.
                          I still don't know how Laithwait lifted that large gyro above his head but however he did it, it wasn't antigravity but probably inertial locking into a gravitational field where the energy to raise it aloft was extracted from the energy in the wheel. No free lunch.
                           
                          David
                           
                           
                           
                           
                           
                           
                           
                           
                           
                          ----- Original Message -----
                          Sent: Friday, October 19, 2012 2:36 AM
                          Subject: [beam] Re: Outer Space Robots WOOT!

                           


                          --- In beam@yahoogroups.com, "David Buckley" <david@...> wrote:
                          >
                          > Connor
                          > If it is the device which falls over at the end in video part 2 he doesn't understand the effects of vibration, friction in the motor bearings or most anything about making stuff. 4 out of 10 for effort, 1 out of 10 for ability, he needs to keep up the art and design classes ;-).
                          > Of course that may not be the video you referred to but without reference links it is hard to tell.
                          > (no link here because I don't want Google finding links to those videos.)
                          >
                          > There are more important points to consider than decay after a few hundred/thousand years. One would be motor life. If it has brushes you are looking at about a few thousand hours brush life maximum. If it has gears you are looking at about the same lifetime. even if it doesn't move a lot the oils in bearings will dry up etc.
                          >
                          > Let's be realistic, if BEAM itself is to survive. let alone the robots. it has to show BEAM robots can do more than stumble a few steps or track the sun or oscillate a bit etc.
                          > So far even including Mark's enthusiastic showing of his walkers we don't have much evidence of that. The circuits have elegance but BEAM robots don't do much. They are not now not competitive with simple program controlled robots - most programs of which are certainly not elegant!
                          > Just as a traditional Beam controller is a physical (electronic gates etc) structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions, so a brain can be built in software which is a virtual structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions.
                          > The principles are the same but the implementation is different. How many neurons could you build in an 8 pin PIC, maybe 5 or 10, lets stick with 5, Then with ten chips (80 pins) you could have 50 neurons. Now how about your cell phone, how many neurons could you build in its processors?
                          > An ordinary house fly has about 1 million neurons. Neurons count when it comes to ability. And even house flies don't survive very long, not even close to the 600 years of Richard Daniel (Clifford D.Simak. All the traps of Earth).
                          > There is nothing horrific about building brains in software. Take biology, start with some atoms and arrange them into molecules, then arrange lots of different types of molecules into cells, then arrange quite a few different types of cells into brains, then inject the electric currents/fields which keep the organism alive, then inject more information - modulations of those electric currents/fields which enable the organism to do something. It is all a hierarchy
                          > At one time building BEAM robots was a good introduction to robotics but that is no longer true, BEAM is on a path to nowhere unless a way is found of scaling up the number and types of neurons and functionality of neuron clusters.
                          > When that is accomplished then maybe work could begin on emulating the Pharaohs with a life of millions of years.
                          > David
                          >

                          YES! Yes yes yes yes yes yes yes! That's exactly the case with BEAM. The problem in my mind is the KISS rule. By that logic, KISS itself, by which almost every consecutive BEAMer follows, is the very path to BEAM's ultimate destruction.

                          However, I don't believe the solution is necessarily in software. Why? Because while superficially software seems to have a very large capacitative advantage over traditional BEAM, which it does, everyone simply fails to make a relative comparison between the two technologies. For example: A typical microcontroller used in small robots, such as that Atmel chip on Solarbotics, contains several million transistors. Now one can virtually construct up to perhaps a 10 neuron system on the chip, having many advantages over hardware built circuits such as convenience, flexibility, reconfigureability, etc. A PC has on average several billion transistors, some of the newer ones up to several trillion transistors. Now how many virtual neurons can that fit onto it? Thousands if not tens or hundreds of thousands of complex interneural pathways and action potentials would be buzzing through that software.

                          Now take physical hardware based neural networks. Traditional BEAM only goes so far as simple coupled oscillators on average, rarely something beyond that, and quite often something borderline to a Braitenburg vehichle. However, what if there was a chip of identical size to a microcontroller and the same number of transistors, but is constructed using a neural architecture rather than a digital one, and each inverter represents a single neuron out of millions of transistors.

                          So I'll let the question sink in before I go: Does the digital architecture really exceed BEAM performance-wise when BEAM is scaled up, or down really, to the same level of complexity? The same goes for scaling down the digital architecture to BEAM's level. In either case. Which system is more capable?

                           

                          P.S. yes that is the video. It may work at a substandard level, but it's only a proof of concept. It's not really supposed to work well quite yet. But it definitely does more than nothing. That's what's important.

                          Connor

                        • Martin McKee
                          Okay, if we are talking, would a BEAM IC be more powerful for reactive robotics applications than software then I think the answer would have to be yes.
                          Message 12 of 23 , Oct 18, 2012
                          • 0 Attachment
                            Okay, if we are talking, would a BEAM IC be more powerful for reactive robotics applications than software then I think the answer would have to be yes.  However, who is going to implement this ( doubtless more efficient and powerful ) IC.  Are we going to be able to convince a team of university researchers that it makes sense? industry? I would doubt it.  Rather, we are going to have to "make due."  Is that a sub-optimal concession? Yes.  Is that BEAMs death knell? Absolutely not.

                            Right now I think that flexibility may be the most important thing for BEAM ( if we want it to advance ).  There are at least two ways that more flexibility could, easily, be made available.  There is the idea of using a CPLD to assist in making a programmable nervous network.  Say this allows for twelve or twenty-four neurons... what could be done with this?  Obviously it is no where near the complexity of the neural network in ANY living creature, however, it is a factor two to four more complex than most "complex" BEAM robots.  Do we know how to design these networks?  I'm not so sure we do.  That is where flexibility ( and programmability ) become important, it allows for lightning fast reconfiguration and testing of different network designs.  Another option is, of course, to use an FPGA for the design instead of a CPLD.  That would give a finer grain architecture and allow for either more complex networks, neurons, or both.  Of course, FPGAs require specific voltage, tight voltage rail control, and lots of power, all things that fit uncomfortably with BEAM philosophy.

                            My idea.  In the interest of full disclosure I will say that I like software... I understand software.  But I am slowly designing a "virtual" BEAM system of my own.  However, it would not just be a simple implementation of nervous networks in software, I have thought about what other features I have wanted.  So, internally there would be several simulated components: virtual Nx neurons ( usable as either Nv or Nu neurons ), summing junctions ( to allow multiple weighted inputs to be routed anywhere ), bias "voltages" ( to allow pulling an input to a fixed value ) and adaptive weights ( or impedances ).  The first three components are basically what is available in standard BEAM ( slightly generalized perhaps ), the last is something different.  The weights could act as something as simple as a resistor... or more complex, a resistor and diode ( for instance a PNC tap ), but the weights are adaptive... they learn.  This would be EVERY weight in the system.  Every weight would act as a type of memory.  And the effect of the memory could be flexible as well, because this is being "wired" inside software it is not difficult to treat every input the same way as any other.  Therefore, outputs from neurons ( or summing points, or analog inputs to the chip ) could be used to modulate the function of these adaptive weights which then control the function of the main neurons.

                            I could write much more, but things are still not quite taking shape.  Here is, generally, what I am thinking.  The first board will have 16 I/Os each pin can be configured either for digital output or analog input.  Inside the chip there would be at least 30 virtual Nx neurons, some number of summing points and some number of bias values, at least 100 adaptive weights.  It will be a struggle, but I believe that an update rate of 250Hz for the network state is possible ( given the selected microcontroller running at 32MHz ).  The system wouldn't update quickly enough to do anything like a servocore, but it should allow for implementation of something like a powersmart architecture, and it is definitely fast enough to make a standard microcore or bicore look like its realtime analog.  And what is the difference in the available processing power? with 30 neurons and 100 memory elements.  Compared to a standard BEAM controller the computational ability is monumental.

                            The real advantage, however, is that the board would be designed to connect to a computer via USB for configuration and debugging.  Computer software that allowed control networks to be designed would also allow them to be immediately uploaded, tested, modified, and stored for later use ( or reference ).  Once it becomes clear that a certain architecture is working well, it would be possible to use that as a template to develop other, yet more complex, designs.  It is incremental steps toward being able to design more capable networks.  Then, when there are implementations ( in software, CPLDs, FPGAs, whatever ) that show the increased capabilities, there is support for actually developing the hardware to do this "the right way."

                            I don't know... with something as potentially complicated ( as opposed to complex... ) as reactive/recurrent neural/nervous networks, I think that the most important thing might just be simplifying the process of searching the solution space... there are a lot of unanswered questions.

                            Martin Jay McKee

                            On Thu, Oct 18, 2012 at 7:36 PM, connor_ramsey@... <connor_ramsey@...> wrote:
                             


                            --- In beam@yahoogroups.com, "David Buckley" <david@...> wrote:
                            >
                            > Connor
                            > If it is the device which falls over at the end in video part 2 he doesn't understand the effects of vibration, friction in the motor bearings or most anything about making stuff. 4 out of 10 for effort, 1 out of 10 for ability, he needs to keep up the art and design classes ;-).
                            > Of course that may not be the video you referred to but without reference links it is hard to tell.
                            > (no link here because I don't want Google finding links to those videos.)
                            >
                            > There are more important points to consider than decay after a few hundred/thousand years. One would be motor life. If it has brushes you are looking at about a few thousand hours brush life maximum. If it has gears you are looking at about the same lifetime. even if it doesn't move a lot the oils in bearings will dry up etc.
                            >
                            > Let's be realistic, if BEAM itself is to survive. let alone the robots. it has to show BEAM robots can do more than stumble a few steps or track the sun or oscillate a bit etc.
                            > So far even including Mark's enthusiastic showing of his walkers we don't have much evidence of that. The circuits have elegance but BEAM robots don't do much. They are not now not competitive with simple program controlled robots - most programs of which are certainly not elegant!
                            > Just as a traditional Beam controller is a physical (electronic gates etc) structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions, so a brain can be built in software which is a virtual structure in which information (what is in the mind of the robot) ebbs and flows and controls the robot's actions.
                            > The principles are the same but the implementation is different. How many neurons could you build in an 8 pin PIC, maybe 5 or 10, lets stick with 5, Then with ten chips (80 pins) you could have 50 neurons. Now how about your cell phone, how many neurons could you build in its processors?
                            > An ordinary house fly has about 1 million neurons. Neurons count when it comes to ability. And even house flies don't survive very long, not even close to the 600 years of Richard Daniel (Clifford D.Simak. All the traps of Earth).
                            > There is nothing horrific about building brains in software. Take biology, start with some atoms and arrange them into molecules, then arrange lots of different types of molecules into cells, then arrange quite a few different types of cells into brains, then inject the electric currents/fields which keep the organism alive, then inject more information - modulations of those electric currents/fields which enable the organism to do something. It is all a hierarchy
                            > At one time building BEAM robots was a good introduction to robotics but that is no longer true, BEAM is on a path to nowhere unless a way is found of scaling up the number and types of neurons and functionality of neuron clusters.
                            > When that is accomplished then maybe work could begin on emulating the Pharaohs with a life of millions of years.
                            > David
                            >

                            YES! Yes yes yes yes yes yes yes! That's exactly the case with BEAM. The problem in my mind is the KISS rule. By that logic, KISS itself, by which almost every consecutive BEAMer follows, is the very path to BEAM's ultimate destruction.

                            However, I don't believe the solution is necessarily in software. Why? Because while superficially software seems to have a very large capacitative advantage over traditional BEAM, which it does, everyone simply fails to make a relative comparison between the two technologies. For example: A typical microcontroller used in small robots, such as that Atmel chip on Solarbotics, contains several million transistors. Now one can virtually construct up to perhaps a 10 neuron system on the chip, having many advantages over hardware built circuits such as convenience, flexibility, reconfigureability, etc. A PC has on average several billion transistors, some of the newer ones up to several trillion transistors. Now how many virtual neurons can that fit onto it? Thousands if not tens or hundreds of thousands of complex interneural pathways and action potentials would be buzzing through that software.

                            Now take physical hardware based neural networks. Traditional BEAM only goes so far as simple coupled oscillators on average, rarely something beyond that, and quite often something borderline to a Braitenburg vehichle. However, what if there was a chip of identical size to a microcontroller and the same number of transistors, but is constructed using a neural architecture rather than a digital one, and each inverter represents a single neuron out of millions of transistors.

                            So I'll let the question sink in before I go: Does the digital architecture really exceed BEAM performance-wise when BEAM is scaled up, or down really, to the same level of complexity? The same goes for scaling down the digital architecture to BEAM's level. In either case. Which system is more capable?

                             

                            P.S. yes that is the video. It may work at a substandard level, but it's only a proof of concept. It's not really supposed to work well quite yet. But it definitely does more than nothing. That's what's important.

                            Connor


                          • G.A. Anderson
                            I m new here, but I would just like to add my $.02. I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous
                            Message 13 of 23 , Oct 19, 2012
                            • 0 Attachment
                              I'm new here, but I would just like to add my $.02.

                              I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                              I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                              Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                              Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                              Guy
                            • David Buckley
                              Guy ... - exactly, but what patterns? That is what nobody seems to know. Overlaying patterns onto a particular set of neurons is not really a good idea because
                              Message 14 of 23 , Oct 19, 2012
                              • 0 Attachment
                                Guy
                                > patterns between the neurons
                                - exactly, but what patterns? That is what nobody seems to know.
                                Overlaying patterns onto a particular set of neurons is not really a good idea because over the years people have gone on about how neural nets could store multiple recognition patters but it just meant the patterns were impossible (at lest very very difficult) to understand, rather like superimposed holograms.
                                 
                                >Some of the greater BEAM robots seem to have minds of their own, with their behaviours
                                > somewhat unpredictable. With a program-controlled robot, all it could really hope to
                                > be is a simulation.
                                That all depends on how the program is written. Programs which set out to be a simulation end up as a simulation. However in software it is possible to construct a virtual machine/brain where the flow of what happens is not written into the program but depends solely on the data the virtual machine works with. Just as the silicon of chips used in BEAM brains doesn't dictate what happens but rather it is the pulsating electricity which does. All that happens when a virtual BEAM brain is built rather than a silicon BEAM brain is that the time granules get bigger. But so what, BEAM robots are not what I would describe as lightning fast, and in any case everything in a wetware brain happens in a granular fashion.
                                 
                                What isn't well understood, or really not understood at all is how the ebb and flow of the information patterns in the silicon-gate-neurons or wetware-neurons makes things alive.
                                What neuronal patterns are good.
                                What information patterns lead to interesting results, rather like what patterns in Conway's Life do interesting and sustaining things.
                                 
                                Help on this would be appreciated by all.
                                 
                                David
                                 
                                 
                                 
                                 
                                 
                                ----- Original Message -----
                                Sent: Friday, October 19, 2012 3:15 PM
                                Subject: [beam] Re: Outer Space Robots WOOT!

                                 

                                I'm new here, but I would just like to add my $.02.

                                I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                Guy

                              • Martin McKee
                                Agreed that the big question is, what is a productive structure for the system? My interest in BEAM is ( and always has been ) that it is a vast
                                Message 15 of 23 , Oct 19, 2012
                                • 0 Attachment
                                  Agreed that the big question is, "what is a productive structure for the system?"  My interest in BEAM is ( and always has been ) that it is a vast simplification of neural network systems.  That means two things.  First, the networks are both designable and explainable.  This is in stark contrast to many of the neural networks in research that are learned against a training set.  The networks solve the problem ( sometimes surprisingly well ) but it is often impossible to explain how they do it.  Similarly, BEAM networks can solve problems in much the way that they are designed, but, at the same time, have surprising emergent behaviors.  The second thing that comes from BEAM's simplifying assumptions, is that it becomes possible to implement the general structure in many more ways.  The standard resistor, capacitor, inverter based Nv neuron can be represented, abstractly, as a filter and threshold... implemented with op-amps, with discretes and the registers of a CPLD, as "objects" in a program, or as pure digital modules in an FPGA.  It allows for many varying paths of research using the same concepts but different "hardware" -- always good in my book.

                                  Martin Jay McKee

                                  On Fri, Oct 19, 2012 at 9:04 AM, David Buckley <david@...> wrote:
                                   

                                  Guy
                                  > patterns between the neurons
                                  - exactly, but what patterns? That is what nobody seems to know.
                                  Overlaying patterns onto a particular set of neurons is not really a good idea because over the years people have gone on about how neural nets could store multiple recognition patters but it just meant the patterns were impossible (at lest very very difficult) to understand, rather like superimposed holograms.
                                   
                                  >Some of the greater BEAM robots seem to have minds of their own, with their behaviours
                                  > somewhat unpredictable. With a program-controlled robot, all it could really hope to
                                  > be is a simulation.
                                  That all depends on how the program is written. Programs which set out to be a simulation end up as a simulation. However in software it is possible to construct a virtual machine/brain where the flow of what happens is not written into the program but depends solely on the data the virtual machine works with. Just as the silicon of chips used in BEAM brains doesn't dictate what happens but rather it is the pulsating electricity which does. All that happens when a virtual BEAM brain is built rather than a silicon BEAM brain is that the time granules get bigger. But so what, BEAM robots are not what I would describe as lightning fast, and in any case everything in a wetware brain happens in a granular fashion.
                                   
                                  What isn't well understood, or really not understood at all is how the ebb and flow of the information patterns in the silicon-gate-neurons or wetware-neurons makes things alive.
                                  What neuronal patterns are good.
                                  What information patterns lead to interesting results, rather like what patterns in Conway's Life do interesting and sustaining things.
                                   
                                  Help on this would be appreciated by all.
                                   
                                  David
                                   
                                   
                                   
                                   
                                   
                                  ----- Original Message -----
                                  Sent: Friday, October 19, 2012 3:15 PM
                                  Subject: [beam] Re: Outer Space Robots WOOT!

                                   

                                  I'm new here, but I would just like to add my $.02.

                                  I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                  I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                  Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                  Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                  Guy


                                • Michael Demaddis
                                  You sure onto something, tiny insects can do so much with so little neurons, its the way they are put together and how they learn including what is already in
                                  Message 16 of 23 , Oct 19, 2012
                                  • 0 Attachment
                                    You sure onto something, tiny insects can do so much with so little neurons, its the way they are put together and how they learn including what is already in there, such as instinct for different creatures, hence lots of animals can walk immediately etc.


                                     
                                    -Michael

                                    From: G.A. Anderson <autisticando@...>
                                    To: beam@yahoogroups.com
                                    Sent: Friday, October 19, 2012 10:15 AM
                                    Subject: [beam] Re: Outer Space Robots WOOT!

                                     
                                    I'm new here, but I would just like to add my $.02.

                                    I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                    I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                    Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                    Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                    Guy


                                  • David Buckley
                                    So little neurons!!! As I said before an ordinary housefly has about 1 million. That s 100,000 times as many as a complicated BEAM robot. And BEAM robots and
                                    Message 17 of 23 , Oct 19, 2012
                                    • 0 Attachment
                                      
                                      So little neurons!!!
                                      As I said before an ordinary housefly has about 1 million.
                                      That's 100,000 times as many as a complicated BEAM robot. And BEAM robots and Processor based robots don't have the thousands of neuronal interconnections and hundreds (thousands) of sensors, so we have a long way to go.
                                       
                                       
                                       
                                      ----- Original Message -----
                                      Sent: Friday, October 19, 2012 5:31 PM
                                      Subject: Re: [beam] Re: Outer Space Robots WOOT!

                                       

                                      You sure onto something, tiny insects can do so much with so little neurons, its the way they are put together and how they learn including what is already in there, such as instinct for different creatures, hence lots of animals can walk immediately etc.


                                       
                                      -Michael

                                      From: G.A. Anderson <autisticando@...>
                                      To: beam@yahoogroups.com
                                      Sent: Friday, October 19, 2012 10:15 AM
                                      Subject: [beam] Re: Outer Space Robots WOOT!

                                       
                                      I'm new here, but I would just like to add my $.02.

                                      I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                      I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                      Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                      Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                      Guy


                                    • Martin McKee
                                      True, and true. But, the question ( or my question anyway ) is how we can most effectively organize the neurons. The approach that most connectionist
                                      Message 18 of 23 , Oct 19, 2012
                                      • 0 Attachment
                                        True, and true.  But, the question ( or my question anyway ) is how we can most effectively organize the neurons.

                                        The approach that most connectionist researchers ( neural network models ) have taken is to use "learning rules" to organize the networks based upon the the data they are presented.  This works surprisingly well, with some large networks ( >100,000 nodes ) being able to compete with humans at handwriting recognition tasks ( and other, similar, pattern matching "stuff" ).  The networks are all but opaque to deciphering how they work though.  Beyond that, the learning procedure is fragile.  It is very easy to construct a data set that leads to networks that perform badly or fail to generalize ( the, so called, overfitting problem ).  

                                        One of the reasons for mapping the neural networks of simple organisms is to gain insight about how to create different systems with the types of neurons available.  That is simplified by finding groups that work together and making modular "cuts" there.  But, in the end, it is a question of complexity.  How much can we understand "at a glance?"

                                        Looking forward, I would think it would be pretty easy to implement simple enough neurons and memory elements that I could fit 100s of neurons and 1000s of connections into a low power microcontroller ( a 32-bit ARM ).  Honestly, the biggest constraint is likely to be the amount of SRAM available on the chip.  If many of these modules could be used on a robot, there could be 1000s of total neurons -- nothing like a house fly ( down by several orders of magnitude ), but, potentially, a huge advance from current BEAM networks.

                                        Martin Jay McKee

                                        On Fri, Oct 19, 2012 at 11:27 AM, David Buckley <david@...> wrote:
                                         

                                        So little neurons!!!
                                        As I said before an ordinary housefly has about 1 million.
                                        That's 100,000 times as many as a complicated BEAM robot. And BEAM robots and Processor based robots don't have the thousands of neuronal interconnections and hundreds (thousands) of sensors, so we have a long way to go.
                                         
                                         
                                         
                                        ----- Original Message -----
                                        Sent: Friday, October 19, 2012 5:31 PM
                                        Subject: Re: [beam] Re: Outer Space Robots WOOT!

                                         

                                        You sure onto something, tiny insects can do so much with so little neurons, its the way they are put together and how they learn including what is already in there, such as instinct for different creatures, hence lots of animals can walk immediately etc.


                                         
                                        -Michael

                                        From: G.A. Anderson <autisticando@...>
                                        To: beam@yahoogroups.com
                                        Sent: Friday, October 19, 2012 10:15 AM
                                        Subject: [beam] Re: Outer Space Robots WOOT!

                                         
                                        I'm new here, but I would just like to add my $.02.

                                        I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                        I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                        Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                        Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                        Guy



                                      • BR
                                        ... The Society of Mind . Minsky, Marvin. Simon and Schuster, New York. March 15, 1988. ISBN 0-671-65713-5
                                        Message 19 of 23 , Oct 19, 2012
                                        • 0 Attachment
                                          On 2012-10-19 8:04 AM, David Buckley wrote:
                                          >
                                          > What isn't well understood, or really not
                                          > understood at all is how the ebb and flow
                                          > of the information patterns in the
                                          > silicon-gate-neurons or wetware-neurons
                                          > makes things alive.

                                          "The Society of Mind". Minsky, Marvin. Simon and Schuster, New York.
                                          March 15, 1988. ISBN 0-671-65713-5
                                        • David Buckley
                                          I possess a copy and have read it. I used to be impressed with Minsky and the fact that when an undergrad he had two labs with budgets, one even in a different
                                          Message 20 of 23 , Oct 20, 2012
                                          • 0 Attachment
                                            I possess a copy and have read it. I used to be impressed with Minsky and the fact that when an undergrad he had two labs with budgets, one even in a different Department.
                                            However does the book help any in moving from a ten neuron robot to a hundred neuron robot to a thousand neuron robot and so on?
                                            What he talks about in the book has been known for centuries. It is why crazy people are called crazy because just like the glaze on antique porcelain breaks into little areas so their minds are broken into little separate areas. Healthy minds have bridges between areas.
                                            It doesn't help much in making brains otherwise somebody surely would have said 'oh I see how to do it now.' and gone away and done it.
                                             
                                             
                                             
                                            ----- Original Message -----
                                            From: BR
                                            Sent: Saturday, October 20, 2012 5:49 AM
                                            Subject: Re: [beam] Re: Outer Space Robots WOOT!

                                             

                                            On 2012-10-19 8:04 AM, David Buckley wrote:
                                            >
                                            > What isn't well understood, or really not
                                            > understood at all is how the ebb and flow
                                            > of the information patterns in the
                                            > silicon-gate-neurons or wetware-neurons
                                            > makes things alive.

                                            "The Society of Mind". Minsky, Marvin. Simon and Schuster, New York.
                                            March 15, 1988. ISBN 0-671-65713-5

                                          • David Buckley
                                            Martin You read my mind! However tonight a friend showed me his new PIC board with a tiny 6 pin smtPIC also available as an 8 pin DIP. It has 2 PWM, 1 A/D,
                                            Message 21 of 23 , Oct 21, 2012
                                            • 0 Attachment
                                              Martin
                                              You read my mind!
                                              However tonight a friend showed me his new PIC board with a tiny 6 pin smtPIC also available as an 8 pin DIP.
                                              It has 2 PWM, 1 A/D, software selectable clock - 31KHz - 32MHz, an 8x8 user configurable ASIC, has 25mA i/o capability and runs at 50microamps.
                                              At present he is writing a macro library for it. No I don't remember the type number!!!
                                              He showed it bidirectionaly controlling a motor (through a H bridge) and current sensing stall at each end of travel to reverse direction. With a split battery supply he could do two motors. The downside is you need a PicKit-3 to program it.
                                              Somewhere between it and the PIC32 it must be possible, taking inspiration from nature, to figure out connection schemes which will allow robots to do much more than stumble forwards and back, and do it in some understandable way which is expandable.
                                              David
                                              .
                                               
                                               
                                               
                                               
                                              ----- Original Message -----
                                              Sent: Friday, October 19, 2012 8:26 PM
                                              Subject: Re: [beam] Re: Outer Space Robots WOOT!

                                               

                                              True, and true.  But, the question ( or my question anyway ) is how we can most effectively organize the neurons.


                                              The approach that most connectionist researchers ( neural network models ) have taken is to use "learning rules" to organize the networks based upon the the data they are presented.  This works surprisingly well, with some large networks ( >100,000 nodes ) being able to compete with humans at handwriting recognition tasks ( and other, similar, pattern matching "stuff" ).  The networks are all but opaque to deciphering how they work though.  Beyond that, the learning procedure is fragile.  It is very easy to construct a data set that leads to networks that perform badly or fail to generalize ( the, so called, overfitting problem ).  

                                              One of the reasons for mapping the neural networks of simple organisms is to gain insight about how to create different systems with the types of neurons available.  That is simplified by finding groups that work together and making modular "cuts" there.  But, in the end, it is a question of complexity.  How much can we understand "at a glance?"

                                              Looking forward, I would think it would be pretty easy to implement simple enough neurons and memory elements that I could fit 100s of neurons and 1000s of connections into a low power microcontroller ( a 32-bit ARM ).  Honestly, the biggest constraint is likely to be the amount of SRAM available on the chip.  If many of these modules could be used on a robot, there could be 1000s of total neurons -- nothing like a house fly ( down by several orders of magnitude ), but, potentially, a huge advance from current BEAM networks.

                                              Martin Jay McKee

                                              On Fri, Oct 19, 2012 at 11:27 AM, David Buckley <david@...> wrote:
                                               

                                              So little neurons!!!
                                              As I said before an ordinary housefly has about 1 million.
                                              That's 100,000 times as many as a complicated BEAM robot. And BEAM robots and Processor based robots don't have the thousands of neuronal interconnections and hundreds (thousands) of sensors, so we have a long way to go.
                                               
                                               
                                               
                                              ----- Original Message -----
                                              Sent: Friday, October 19, 2012 5:31 PM
                                              Subject: Re: [beam] Re: Outer Space Robots WOOT!

                                               

                                              You sure onto something, tiny insects can do so much with so little neurons, its the way they are put together and how they learn including what is already in there, such as instinct for different creatures, hence lots of animals can walk immediately etc.


                                               
                                              -Michael

                                              From: G.A. Anderson <autisticando@...>
                                              To: beam@yahoogroups.com
                                              Sent: Friday, October 19, 2012 10:15 AM
                                              Subject: [beam] Re: Outer Space Robots WOOT!

                                               
                                              I'm new here, but I would just like to add my $.02.

                                              I have a M.A. in Experimental Psychology, and as such I know a bit about how the brain is. Not a tremendous amount, but I still know more than most.

                                              I don't think that it is the amount of neurons in gross that will determine the complexity and survivability of a BEAM robot, but rather the patterns between the neurons. This is good news all around because complex BEAM robots do not necessarily need to be a large as we once previously thought (or at least this is possible). Right now, the circuit components that make up the BEAM robot's "brain" are much larger than virtually any neuron, but this may change in due time. As such, we should focus on constructing the greatest amount of patterns with a relatively lesser number of components.

                                              Another thing is that a BEAM robot has one innate advantage to a program-controlled one: it is much more similar to something that is actually alive. Some of the greater BEAM robots seem to have minds of their own, with their behaviors somewhat unpredictable. With a program-controlled robot, all it could really hope to be is a simulation. Programs are really just that: simulations. I could go on more about it, but I'm running out of energy.

                                              Anyways I hope that I said some insightful things. I can barely even take motors out of CD players so obviously I lack experience, but I nonetheless think that I am onto something.

                                              Guy



                                            • connor_ramsey@ymail.com
                                              Well, all this chatter has got me thinking. Don t they use programs now to automatically construct the pathways in a computer chip instead of hand-written
                                              Message 22 of 23 , Oct 21, 2012
                                              • 0 Attachment

                                                Well, all this chatter has got me thinking. Don't they use programs now to automatically construct the pathways in a computer chip instead of hand-written scematics? And then that same machine goes ahead and prints the IC. Surely the same could easily be done for the hypothetical "BEAM IC". Although this could only be accomplished once we do understand the types of interneural connections that work best. But like I've said before, we can always use the schematic of C. elegas' neural network. All 302 of the worm's neurons and all the individual ganglion structures are described and displayed in detail at wormatlas.com, even delving into the specific function of each consecutive neuron and interneural connections! Trust me here, if there's anywhere that we can figure out the successful patterns to implement, this is absolutely it.

                                                Go ahead, take a look! Connor.

                                                Oh, and if your going to do that, Richard, might I suggest using an H-switch network to operate the CPLD's virtual connections?

                                                 

                                              • Martin McKee
                                                Designing a BEAM IC isn t the problem ( assuming we can figure out how to connect neurons that is ), it is getting them manufactured -- big bucks! It is
                                                Message 23 of 23 , Oct 21, 2012
                                                • 0 Attachment
                                                  Designing a "BEAM IC" isn't the problem ( assuming we can figure out how to connect neurons that is ), it is getting them manufactured -- big bucks!  It is certainly possible ( and often done ) to generate some of the HDL ( Hardware Description Language ) for an IC, but there is more to it than simply designing the circuit.  Once the code ( some of which will be produced by hand ) is complete, it is necessary to map the design to the silicon ( a, typically, long and expensive process ), the "process" must be laid out ( ICs are built in layers... if something is done out of order, it won't work ).  Then masks are made for photolithography ( expensive again for small processes nodes ) and a first batch of chips can be run.  Some of this can be trimmed if the chip were implemented in an ASIC process  which, typically, works by burning out fuses, but you're still talking about >$50,000 for a first chip run ( unless, of course, you happen to have access to your own chip fab... which some universities do ).  ASIC won't work if we want the "BEAM IC" to be analog, so look at ( at least ) tripling that estimate for a first run...  Hopefully the first run of silicon would work perfectly... typically that's not the case.  Semiconductor design and manufacture is both expensive and time consuming.  BEAM simply doesn't have the clout at this point to interest anyone with the resources.  That is why so few devices on the market have moved from custom chips to microcontrollers, microprocessors or programmable logic ( CPLD or FPGA ); it is more financially viable... even for major manufacturers.

                                                  Certainly, the H-Switch network has possibilities... but then, isn't that part of our problem? so many potentially useful components with an ( almost ) infinite number of ways to combine them?  I am considering including an iMx style component in my virtual BEAM "program" ( though, the program only would provide the framework, a configuration defines the functionality ).  I've considered other components as well.  Most of my focus has been on generalizing the functionality.  For instance, my multiplexer ( as with all of the components ) would actually function based entirely upon analog signals, and the switch would, actually, be a mixer.  The "switch" input would actually designate the mixture ratio.  An input of -1 would send the inputs to opposite outputs, switch of 1 would send the inputs to the "normal" outputs, a switch of 0 would send the average of each input to both outputs, and so on.  Are there other possible connection methods? I can think of several: a grid with "inputs" on one axis and "outputs" on the other with programmable connections, the "output" could be the average of the inputs ( for analog functionality ) or some logic function ( AND, OR, etc. ) thereof, It would be possible to have specific sum junctions with programmable input weights and digital values, it is also possible to separate out the sum junction and weights as separate components ( what I plan to do ).

                                                  I seem to be doing quite a bit of theorizing lately, and I'm doing very little "doing."  I think I need to attempt to remedy that situation.  I'm going to try to get something I can play with ( on the computer ) by the end of the week... even if it's just some subset of the components.  One of the ideas I've had is that, by doing everything in software anyway, I will be able to run the networks in a fully simulated environment.  The major downside of this is, of course, that, as Rodney Brooks has stated, "The world is it's own best model."  By simulating a system entirely in the computer, it will get rid of some of the complexities of "real life."  On the other hand, getting rid of complexity can be a good thing at times.  Also, it would mean that I get to test the functionality of my algorithms in a known environment.  I think I have two specific tasks to work toward.  See if my current set of components are sufficient to implement an H-Switch type component, and implement a basic bicore in the virtual neurons.

                                                  Martin Jay McKee

                                                  P.S. I am likely to continue using the term virtual neuron as opposed to simulated neuron for the very simple reason that, from outside the chip, they both look the same.  If the virtual ( or "simulated" ) neuron affects a certain output as a result of a specific input, it is doing a very real job.  In testing on a computer ( where stimuli are fabricated as well ) I don't really have any issue with the term simulated neuron.

                                                  On Sun, Oct 21, 2012 at 8:02 PM, connor_ramsey@... <connor_ramsey@...> wrote:
                                                   

                                                  Well, all this chatter has got me thinking. Don't they use programs now to automatically construct the pathways in a computer chip instead of hand-written scematics? And then that same machine goes ahead and prints the IC. Surely the same could easily be done for the hypothetical "BEAM IC". Although this could only be accomplished once we do understand the types of interneural connections that work best. But like I've said before, we can always use the schematic of C. elegas' neural network. All 302 of the worm's neurons and all the individual ganglion structures are described and displayed in detail at wormatlas.com, even delving into the specific function of each consecutive neuron and interneural connections! Trust me here, if there's anywhere that we can figure out the successful patterns to implement, this is absolutely it.

                                                  Go ahead, take a look! Connor.

                                                  Oh, and if your going to do that, Richard, might I suggest using an H-switch network to operate the CPLD's virtual connections?

                                                   


                                                Your message has been successfully submitted and would be delivered to recipients shortly.