Another advantage to swarms - graceful degradation
on
Of Ants and Robots
·
· Score: 2, Interesting
The researcher in the project neglected to mention what I see as a huge advantage to using swarm-intelligence: graceful degradation. That is, in most common machines and software, if a single part breaks down, the effects on the systems functioning can be catastrophic. However in other systems, such as a neural network for example, the deterioration of a section of the system will not lead to a total loss of function -the sytem will degrade gracefully. It seems to me a reasonable assumption that this will apply similarly to swarms, so long as there aren't key robots whose existence is essential to the proper functioning of the system.
Re:Good example of emergent behavior
on
Of Ants and Robots
·
· Score: 2, Informative
I disagree with your refutation that this is an example of emergent behaviour; I believe this is just the type of case where we would want to use the term "emergence". The behaviour that is being described as emergent is the ability of the units to work together to exhibit some sort of collectively intelligent behaviour, as if they were being guided by some sort of overseer. However, the lack of existence of any such overseer leads us to call this behaviour emergent -it emerges as a result of the individual units and the way they interact.
This is precisely the same way we use the term "emergence" when talking about neural nets. There are very simple artificial neurons, and yet the behaviour of the system can seem quite advanced. This behaviour is in no way programmed into the nodes themselves -it emerges as a result of the way they interact.
The fact that there is a "collective idea" about how to solve the problem doesn't mean it is not emergent behaviour, because we could say the same thing about neural nets. If that were the case, then there would be no need for the term "emergence" at all. However, this wouldn't seem right at all, since we don't really have any better way to describe what happens when a collection of input/ouput nodes is able to somehow accurately predict English past tense (among other nifty things neural nets can do).
Your statement about how chemical trails don't "explain their ability to work out the logistics on the fly". Hits directly on this issue. Of course the ants can't work out the logistics of the problem -they couldn't even understand it if the answer were given to them. However, the collective group of ants can be uncontroversially described as "working out" the logistics of the problem. This should strike you as odd, and is exactly why we would want to talk about emergent behaviour in this case. The same thought can be applied to the complex structures that the army ants can build.
At some time in the future this will be a widely debated and incredibly important issue. That time will come as scientists in AI come closer and closer to producing a robot that behaves as a free-thinking and rational being, with independent thoughts and feelings (if that is indeed possible)
The question that comes up is whether they are actually feeling, thinking, and conscious beings, or whether they are just imitations. The answer is yes, they are. Why? The physical form of the being doesn't matter, be it neurons and chemicals, or circuit boards and wiring.
Consider the following thought experiment: Imagine the neurons in your brain suddenly lost their ability to fire. Unfortunately, this means you wouldn't have thoughts or consciousness. Fortunately, at precisely the same time, a magical microscopic mind-gnome appears inside your brain, and he wants to help. Every single time one of your neurons would be firing, he hops over there at incredible speeds and gives that neuron a shove. Effectively, your brain functions exactly the same way, although it has changed physical form. Now ask yourself, under such situations, would you consider yourself a conscious being? Of course you would!
(Haugland's thought experiment "neuron tickler")
In the thought experiment above, you can change other parts of the brain and still the function of the brain is still the same: it produces thoughts and consciousness. Replace the brain with circuit boards and the proper AI programming. You now have a thinking, conscious robot.
There is nothing that even suggests that a robot who behaves like we do, thinks like we do, and feels like we is not a possibility. Every day researchers get closer to the day when a robot poet will pick up a pen and write his love poem, or a robot philosopher will ponder whether or not it has a soul, or if such a thing even exists. And that brings us to a difficult part of the robot rights debate-the notion of a soul.
The soul argument for why robots should not have rights has three premises. The first of course is that souls exist. The second is that only beings with a soul are given rights. The third is that robots do not have souls. A robot philosopher defending his rights would most likely opt to argue the third is not true, an atheist might go after the first. But, we can be certain that if we are to convince someone who believes in the existence of a soul that robots deserve rights, then we must accept the notion that souls exist. It is from this standpoint that I will show why, even if souls exist, that robots should be according rights.
First of all, does a soul effect our functioning in any way? Does a soul change how we think, act, or behave? All modern evidence points to no. A human does everything it does with what is there physically. Neurons, chemicals, and electrical signals govern us entirely. If (as it appears so far) a soul has no effect on our functioning, there is no way to know that we even have souls at all.
How do we know we have a soul (or why do we think we have a soul)? Because we laugh, cry, love, hate, think, feel, and most importantly, we sense something special within ourselves, something that is not explainable by science. Imagine a robot that felt all the same things? Would we not say that they too had a soul? Some might argue that only humans can have souls, but is that logical, and is that fair? Of course it is neither.
There is of course the possibility that a soul is required to function as a human. Perhaps no matter how much research goes into the field of AI, we will never create a robot that behaves as humans do, because it lacks the essential component: the soul. In such a case we will be assured through scientific means of the existence of a soul, and our inability to create one. However, if we are able to create a robot which functions as we do, and a soul is required for functioning, then that robot must have a soul.
Is it impossible for a robot to have a soul? No. Consider the following thought experi
minimum mass of Krypton!
on
Comic Book Physics
·
· Score: 5, Interesting
Pieces of kryptonite are found scattered about the planet. We are told these are shards from the destruction of Krypton. If we assume that the shards were evenly distributed in all directions, we can determine the minimum size the planet Krypton must have been.
First off we need the distance. Let's assume Krypton circled the nearest star to our own (we are looking for the minimum size of Krypton). Proxima Centauri (or Alpha Centauri C) is only 4.22 light-years away. (393 927 289 812km)
Imagine a sphere whose radius extends from where the planet Krypton used to be, to the earth. The surface area of this sphere represents the 3-d area across which the shards of kryptonite were distributed. This sphere has a surface area of 4.87508x10^23km(standard calculation).
The earths radius represents a fraction of this total surface area. The earths radius is 6.3781 x 10^3km. multiply by pi to get the area (the area is 2-d -ie not squared- because the surface of a sphere is 2-d). The next step is comparing this 2-d surface area to the surface area of the imaginary sphere we got above. The result: the earth represents a TINY 4.110086 x 10^-18% of the surface area of our Krypton-explosion sphere. If we multiply the amount of kryptonite on earth by the inverse of this number, we get the amount of Krypton that is scattered around the entire surface area of the sphere.
And how much kryptonite is on the Earth? damned if I know, so let's just estimate based on what we know of the series. It's been made into various weapons and devices, been sold over the blackmarket, been hidden in secret storage areas, been acquired by every evil organization or villian ever, so presumably the amount on Earth is quite high. BUT, we are calculating for a minimum size of Krypton, so we'll estimate low. 10kg seems more than fair. Now, 2/3 the Earths surface is water, and i haven't heard of any kryptonite being recovered from undersea explorations, so that 10kg found on earth was the 1/3rd that hit the land. So, 30kg hit the Earth. Also consider burning up on reentry. I don't know of kryptonite being indestructible, and it has been made into a liquid at least once in Superman history. Its Probable that at least 90% was burned up in reentry. (If someone with more precise figures and re-do calcs t'would be appreciated). so, the 30kg that hit the earth represents only 10% of the 300kg that hit the atmosphere.
multiply this by the inverse of this by the inverse of the fraction that represents the surface of our Krypton-explosion sphere over our earths surface area sphere. The result: The planet Krypton weighed an absolute minimum of 7.299x10^19kg. By comparison, our sun weighs 2x10^30kg.
Argh. What do you mean "you CAN'T deflect a beam of light with just "Mind Power!""?
Jean Grey typically uses her mind powers to manipulate objects. Presumably, this means she exerts some sort of force on the objects, probably a force similar to gravity. Now the question you have to ask yourself is, can gravity affect light? Is there any reason to suspect that her powers wouldn't be able to affect light as well? It would take alot of focus or 'mind-energy', i agree, but moving objects versus moving light with the power of your mind alone are in principle the same. The only difference between the two is the amount of energy required.
So, could Jean Grey deflect a beam of light? No way, that would require SUPER-POWERS or something...
Slashdot Mirror
The researcher in the project neglected to mention what I see as a huge advantage to using swarm-intelligence: graceful degradation. That is, in most common machines and software, if a single part breaks down, the effects on the systems functioning can be catastrophic. However in other systems, such as a neural network for example, the deterioration of a section of the system will not lead to a total loss of function -the sytem will degrade gracefully. It seems to me a reasonable assumption that this will apply similarly to swarms, so long as there aren't key robots whose existence is essential to the proper functioning of the system.
I disagree with your refutation that this is an example of emergent behaviour; I believe this is just the type of case where we would want to use the term "emergence". The behaviour that is being described as emergent is the ability of the units to work together to exhibit some sort of collectively intelligent behaviour, as if they were being guided by some sort of overseer. However, the lack of existence of any such overseer leads us to call this behaviour emergent -it emerges as a result of the individual units and the way they interact. This is precisely the same way we use the term "emergence" when talking about neural nets. There are very simple artificial neurons, and yet the behaviour of the system can seem quite advanced. This behaviour is in no way programmed into the nodes themselves -it emerges as a result of the way they interact. The fact that there is a "collective idea" about how to solve the problem doesn't mean it is not emergent behaviour, because we could say the same thing about neural nets. If that were the case, then there would be no need for the term "emergence" at all. However, this wouldn't seem right at all, since we don't really have any better way to describe what happens when a collection of input/ouput nodes is able to somehow accurately predict English past tense (among other nifty things neural nets can do). Your statement about how chemical trails don't "explain their ability to work out the logistics on the fly". Hits directly on this issue. Of course the ants can't work out the logistics of the problem -they couldn't even understand it if the answer were given to them. However, the collective group of ants can be uncontroversially described as "working out" the logistics of the problem. This should strike you as odd, and is exactly why we would want to talk about emergent behaviour in this case. The same thought can be applied to the complex structures that the army ants can build.
At some time in the future this will be a widely debated and incredibly important issue. That time will come as scientists in AI come closer and closer to producing a robot that behaves as a free-thinking and rational being, with independent thoughts and feelings (if that is indeed possible)
The question that comes up is whether they are actually feeling, thinking, and conscious beings, or whether they are just imitations. The answer is yes, they are. Why? The physical form of the being doesn't matter, be it neurons and chemicals, or circuit boards and wiring.
Consider the following thought experiment: Imagine the neurons in your brain suddenly lost their ability to fire. Unfortunately, this means you wouldn't have thoughts or consciousness. Fortunately, at precisely the same time, a magical microscopic mind-gnome appears inside your brain, and he wants to help. Every single time one of your neurons would be firing, he hops over there at incredible speeds and gives that neuron a shove. Effectively, your brain functions exactly the same way, although it has changed physical form. Now ask yourself, under such situations, would you consider yourself a conscious being? Of course you would! (Haugland's thought experiment "neuron tickler")
In the thought experiment above, you can change other parts of the brain and still the function of the brain is still the same: it produces thoughts and consciousness. Replace the brain with circuit boards and the proper AI programming. You now have a thinking, conscious robot.
There is nothing that even suggests that a robot who behaves like we do, thinks like we do, and feels like we is not a possibility. Every day researchers get closer to the day when a robot poet will pick up a pen and write his love poem, or a robot philosopher will ponder whether or not it has a soul, or if such a thing even exists. And that brings us to a difficult part of the robot rights debate-the notion of a soul.
The soul argument for why robots should not have rights has three premises. The first of course is that souls exist. The second is that only beings with a soul are given rights. The third is that robots do not have souls. A robot philosopher defending his rights would most likely opt to argue the third is not true, an atheist might go after the first. But, we can be certain that if we are to convince someone who believes in the existence of a soul that robots deserve rights, then we must accept the notion that souls exist. It is from this standpoint that I will show why, even if souls exist, that robots should be according rights.
First of all, does a soul effect our functioning in any way? Does a soul change how we think, act, or behave? All modern evidence points to no. A human does everything it does with what is there physically. Neurons, chemicals, and electrical signals govern us entirely. If (as it appears so far) a soul has no effect on our functioning, there is no way to know that we even have souls at all.
How do we know we have a soul (or why do we think we have a soul)? Because we laugh, cry, love, hate, think, feel, and most importantly, we sense something special within ourselves, something that is not explainable by science. Imagine a robot that felt all the same things? Would we not say that they too had a soul? Some might argue that only humans can have souls, but is that logical, and is that fair? Of course it is neither.
There is of course the possibility that a soul is required to function as a human. Perhaps no matter how much research goes into the field of AI, we will never create a robot that behaves as humans do, because it lacks the essential component: the soul. In such a case we will be assured through scientific means of the existence of a soul, and our inability to create one. However, if we are able to create a robot which functions as we do, and a soul is required for functioning, then that robot must have a soul.
Is it impossible for a robot to have a soul? No. Consider the following thought experi
First off we need the distance. Let's assume Krypton circled the nearest star to our own (we are looking for the minimum size of Krypton). Proxima Centauri (or Alpha Centauri C) is only 4.22 light-years away. (393 927 289 812km)
Imagine a sphere whose radius extends from where the planet Krypton used to be, to the earth. The surface area of this sphere represents the 3-d area across which the shards of kryptonite were distributed. This sphere has a surface area of 4.87508x10^23km(standard calculation).
The earths radius represents a fraction of this total surface area. The earths radius is 6.3781 x 10^3km. multiply by pi to get the area (the area is 2-d -ie not squared- because the surface of a sphere is 2-d). The next step is comparing this 2-d surface area to the surface area of the imaginary sphere we got above. The result: the earth represents a TINY 4.110086 x 10^-18% of the surface area of our Krypton-explosion sphere. If we multiply the amount of kryptonite on earth by the inverse of this number, we get the amount of Krypton that is scattered around the entire surface area of the sphere.
And how much kryptonite is on the Earth? damned if I know, so let's just estimate based on what we know of the series. It's been made into various weapons and devices, been sold over the blackmarket, been hidden in secret storage areas, been acquired by every evil organization or villian ever, so presumably the amount on Earth is quite high. BUT, we are calculating for a minimum size of Krypton, so we'll estimate low. 10kg seems more than fair. Now, 2/3 the Earths surface is water, and i haven't heard of any kryptonite being recovered from undersea explorations, so that 10kg found on earth was the 1/3rd that hit the land. So, 30kg hit the Earth. Also consider burning up on reentry. I don't know of kryptonite being indestructible, and it has been made into a liquid at least once in Superman history. Its Probable that at least 90% was burned up in reentry. (If someone with more precise figures and re-do calcs t'would be appreciated). so, the 30kg that hit the earth represents only 10% of the 300kg that hit the atmosphere.
multiply this by the inverse of this by the inverse of the fraction that represents the surface of our Krypton-explosion sphere over our earths surface area sphere. The result: The planet Krypton weighed an absolute minimum of 7.299x10^19kg. By comparison, our sun weighs 2x10^30kg.
Argh. What do you mean "you CAN'T deflect a beam of light with just "Mind Power!""? Jean Grey typically uses her mind powers to manipulate objects. Presumably, this means she exerts some sort of force on the objects, probably a force similar to gravity. Now the question you have to ask yourself is, can gravity affect light? Is there any reason to suspect that her powers wouldn't be able to affect light as well? It would take alot of focus or 'mind-energy', i agree, but moving objects versus moving light with the power of your mind alone are in principle the same. The only difference between the two is the amount of energy required. So, could Jean Grey deflect a beam of light? No way, that would require SUPER-POWERS or something...