Slashdot Mirror
Working Around Patents with Evolutionary Design
An anonymous reader writes "Using computational trial-and-error allowed a Stanford team to come up with a patent-free WiFi antenna. Patent rules are tricky to formulate as self-interest dictates that the claim is as general as possible. Patent fences effectively can build a substantive competitive barrier to markets. Using evolutionary tactics may be a way to legally and ethically bypass these roadblocks."
and then you patent the resulting design.
Intelligent design loses out yet again.
If X is the new Y, and Y is "X is the new Y", solve for X.
Its time to "fix" this problem by removing software and business methods from the purview of the patent office.
What's to stop the company which is engaging in the exclusionary patenting from running exactly the same algorithm and patenting every viable permutation of "thing X" it "evolves"?
:p
I still think improvement in the patent system still has to be made on the level of scoping patentability, in the long run.
BTW, I accelerated the production of this post by using Intelligent Design instead.
~ Whence do you come, slayer of men, or where are you going, conqueror of space?
But who's to stop the person who wrote the algorithm to patent the solution that bypassed the original patent? Or the algorithm itself for that matter?
Efficient antennas 'designed' by evolution are already in use on spacecraft.
Patents were supposed to encourage innovation, but modern patent law has evolved in a way that makes it more of a hindrance than a help. You basically have to have a large corporation and a battery of lawyers behind you to support your patent application, and the corporations aren't even interested unless they are very sure they can see a path to big profits. For the corporations the big attraction is that the patent grants them monopoly profits, and they could not care less about the social values (or harms) of the innovations themselves. From that purely monetary perspective it makes perfect sense to focus on the value of patents for blocking competitors and for lawsuits--though SCO showed that the strategy doesn't always work.
I think the fundamental problem is that the values of patents are too highly variable, and this variability has completely overwhelmed the simple-minded idea of a temporary monopoly. There are cases where it makes sense to motivate innovation by the exclusive monopoly, but almost never for the specific period of time that is hard-coded into patent law. Some patents should lapse more quickly, though of course the companies will argue they should last *MUCH* longer, and they have a lot of lobbying money to push with. Some patentable ideas are very quick and inexpensive to develop, while others take years and lots of money, but patent law doesn't really consider such trivia.
The bottom line dynamic is that most innovation has to start within an individual, but patents have become a team sport. If you aren't on the right team, it doesn't really matter how innovative your ideas are. You're very unlikely to succeed at the patent game without such a team.
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
Hardly "evolutionary", designing a system that designs trial and error is hardly "evolutonary", its basically an intelligent search of a search space compared against a pattern. Evolution is blind, it has no end goal.
While engineers are not actively designing the product, their jobs are still secure as the companies will always need someone to design the algorithms and to study the product goal to know what parameters to set for the algorithms.
The second concern is irrelevant, as engineers would still follow through with rigorous product testing to determine all the possible outcomes of the design, thus avoiding any unplanned problems.
The great value of these algorithms in producing ten to a hundred fold more efficient, faster, or longer lasting products will guarentee that their use will increase, creating a great demand for engineers who can work with them and who can expect high salary bonuses for such amazing results.
We are made wise not by the collection of our past, but by the responsibility for our future. -George Bernard Shaw
I've just submitted a patent for bypassing patents throuigh evolutionary generation of computer applications...
Of course - it's sometimes hard to decide the amount of effort put into a design, but in general - the scale of invention is ranging from obvious to ground-breaking. In the area of antennas it's not really ground-breaking, and often but not always close to obvious. It's more a question of evolution of a previous design.
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
A friend of mine once told me that this is actually an intended result of patents. Note that a patent applies to a specific way of arriving at something, not the something itself. So, the idea is that if the something is desirable, others will go out of their way to find alternative ways to arrive at something. Some of these might be better than the original. Or new somethings may be encountered along the way (inventions tend to happen by accident, yada yada). Whatever the case, patents foster innovation...in this case, by shutting the door on using what is already known to work.
Please correct me if I got my facts wrong.
"Using computational trial-and-error allowed a Stanford team to come up with a patent-free WiFi antenna. Patent rules are tricky to formulate as self-interest dictates that the claim is as general as possible. Patent fences effectively can build a substantive competitive barrier to markets. Using evolutionary tactics may be a way to legally and ethically bypass these roadblocks."
Two problems:
1. For the past 10+ years I keep seeing various articles talking about evolution design and they are all about antennas and simple analogue circuit designs. Antennas are certainly susceptible to evolutionary design, but if we'll be driving the industry forward we'll need to throw lots of R&D to develop evolutionary design algos that can design something more complex. My point is, it's hugely promising, but it's still not here in a big way.
2. The bigger problem, and which is what caused my exclamation in the title: there's no way to avoid overly broad patents. Evolutionary designs in fact often arrive at designs that match exactly various patents. Which means, when your super computer arrives at a working design, you still need to go through all the tedious work of verifying it's not patented, and if it is, start the algo again and hope for the best.
And the limit for rerunning the algo plenty of times to get patent-free design is the same such as manual design: we don't have infinite time, and the solutions to a problem are sometimes finite, and not that many.
I think patents should be left in place, but their running period should be shortened. The industry is developing at such an amazing pace that we make more progress in an year, than what took 10 years before. The original lawmakers never intended their law to run unmodified in such circumstances.
shhht.... dont give em ideas, the patent holders will use evolutionary algorithms themselves in their next patents to make them ever broader
I would say that this shows quite clearly that software patents are more of an obstruction than anything else, quite contrary to the usual argument that patents are for the advancement of technology.
> A century and a half after Darwin suggested natural selection as the mechanism of evolution, engineers have proved him right once again.
I would challenge the assertion that entering the design parameters and working out which is the best result isn't proof of the origin of the species suggested by Darwin.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
Using this method we found this
http://critticall.com/ArtificialSort.html
Makes a lot of people quite angry and nervous, but the real question is - does it work?
It works fine, thank you!
- Thomas
Perhaps they'll be blocked by Koza's patents on genetic programming.
... until someone patents evolutionary design.
Just means the orginal patent they were trying to circumvent wasnt drawn up properly.
---- Booth was a patriot ----
Patent law has not "evolved". It has been maliciously twisted and distorted by corporate interests. That is a very different thing.
And if you want it to stay anywhere near halfway sane, write your Senators and tell them to vote against their new "patent reform" bill. That would change the law to award patents to the first who apply for a patent, rather than the first to invent. Talk about stifling innovation! That would give all the advantages to corporate lawyers, and our patent system would fail completely in its purpose.
We need to get our patent system working like it used to. And do prevent it from degrading even further, write your Senators and tell them to vote against their "patent reform" bill.
If this became law, it would award patents to the first person who filed for the patent, rather than the inventor. This is such a travesty that I cannot believe that it even passed the House... but it did. If that were to pass, you could say goodbye to innovation in the United States. The corporate lawyers would be able to patent almost anything they wanted, even if they did not invent it. The lowly inventor (who usually does not have the resources to patent immediately) would be cut out of the loop. Goodbye innovation!
It is imperative that this bill not pass. Take the time to write. You will be doing yourselves a favor.
Genetic algorithms are the Ruby on Rails of optimization problems. If you can define a state space and a fitness function, you're almost certainly better off using (non-)linear programming, constraint programming, or a local search method like tabu search.
To quote the AI Bible (AIMA 2e, Russell and Norvig): [It] is not clear whether the appeal of genetic algorithms arises from their performance or from their aesthetically pleasing origins in the theory of evolution.
But because GAs are so intuitive for anyone who's taken high school bio, you get fluff pieces like this one...
Excuse me, but it IS a bad thing, from your point of view, if idiots withing your own society are CAUSING it to "devolve" and be anti-competitive.
Patent trolls usually patent general quite obvious things. GP tend to evolve actually innovative things. If they did it, they would get some good designs rather than the very general noninnovative designs that qualifies them as trolls. However, it is quite true that after getting a good design via GP you can patent it.
It is no longer uncommon to be uncommon.
I had already acknowledged that it depends on your point of view. I do not disagree with you, I just thought this was too obvious to mention.
It could well be, that the University's research costed more, than whatever the patents-holder(s) would've charged for the licensing...
In Soviet Washington the swamp drains you.
In a pig's eye, but thank you for making my point. Meta-moderation is *NEVER* off topic as long as /. asks about it in *EVERY* thread and and after *EVERY* posting. Of course YOU (the troll who marked it as off-topic) like the system. You're one of the BOFHs who is doing so much to make /. so worthless.
Freedom = (Meaningful - Coerced) Choice != (Speech | Beer^2), and sad sock puppets' bad mods avail them naught.
What would happen if open source organizations, such as the FSF, looked at a technology, predicted where it was going, and then came up with the most probable solutions, and patented them?
And then offered indemnity only to those companies which published their source under an open source license?
I hear a lot of whining and moaning about bad patents, but not a lot of people are willing to invest the mere $600 it takes to file a patent application. When you think about it, if even 1/10th of the open source contributors filed patent applications for their inventions in code, Corporate America wouldn't be able to make an electric toothbrush without releasing the source code.
Yet, for all of the collective intelligence of the geek community, we'd rather sit on our collective asses and whine, rather than actually do something about the abuses of patent law.
Don't like patent law? Then get thinking, and at least publish your ideas, if not patent them. The FSF used copyright law to protect the freedom of software; we can use patent law in a similar manner.
The society for a thought-free internet welcomes you.
Is probably one of the best ideas I've heard in a while. It defeats patents, and makes our stuff better. Down with the system!
Hardly "evolutionary", designing a system that designs trial and error is hardly "evolutonary"
That is what evolution is.
its basically an intelligent search of a search space compared against a pattern.
Again, that is what evolution is.
Evolution is blind, it has no end goal.
Evolution improves fitness, just like it does here.
Antenna designs are straightforward physical design patents, and have nothing to do with software or business method patents. Such off the cuff, mis-aimed ranting makes the rest of us who dislike software patents look foolish: please don't do it again.
Did you know that Darwin's book was called not 'The Origin of the Species' but 'The Origin of Species' (or 'On the Origin of Species by [blah blah blah]')? I don't think the book is about what you imagine it to be about, but then again, nor do I think the word 'proof' here is used in the way you imagine it to be used. Ironically enough, however, evolution (which I suspect does not mean what you imagine it to mean) is in fact a theorem, not merely a theory, and (although that is not what was meant here) is actually entirely susceptible to mathematical proof. It has the same kind of logical status as the calculus—which is why it is possible to apply it to software construction, and not just to making babies.
It is, however, logically possible for a theist to doubt evolution even after understanding it, though it does require the generation of quite perverse theories. The simplest I've been able to construct is perhaps that whenever evolution is about to happen, under the normal operation of mathematics and physical laws, a miracle occurs to set everything back the way it was. Further miracles are then of course required to make it look exactly as if the first miracle hadn't occurred, since the overall result needs to be identical to the evolutionary case, but at least the supposed theological requirements are met. (I believe this theological design pattern already appears in Roman Catholic doctrine regarding the scientific measurement of the holy sacrament, subsequent to transubstantiation, so clearly there are those to whom this should not appear to be transparent rubbish.)
All told, the hypothesis that the author of the book of Genesis was a smart poet and not a stupid biologist seems simpler, though, and less demeaning to the deity....
From: http://ic.arc.nasa.gov/projects/esg/publications/lohn_papers/gptp2004.pdf
Figure 1.6. Genotype for evolved antenna ST5-3-10.
The lameness filter is obstacling my path, so I'm tootling my google notbook page with vigor.
http://www.google.com/notebook/fullpage?hl=en#b=BDRqeIgoQi4a_oq4i
-theGreater.
While it is true that evolutionary algorithms can be used to do an end-run around patents, it can also be used for new patents. In my opinion, this is what Genetic Programming, Inc. hopes to do: generate patents through evolutionary algorithms. (See, here, for example, on genetically derived patents -- in section 2). Dr. John Koza, as far as I know, teaches at Stanford and also runs this company, and he is considered the father of genetic programming. He has patents on genetic programming as well.
As far as I know, he has never explicitly given permission for any of his patents to be used license-free commercially or noncommercially.
The reason I have used "as far as I know" and "as far as I know" is that this should be an adequate defense of claims against libel. It is my own opinion and belief that Dr. Koza would not stoop to actually suing based on negative posts such as this one.
--Rob
Towards the Singularity.
> "Antenna designs are straightforward physical design patents, and have nothing to do with software or business method patents. Such off the cuff, mis-aimed ranting makes the rest of us who dislike software patents look foolish: please don't do it again."
Guess you missed the whole "pringles can antenna" debate. According to your argument, someone should now be able to patent using a pringles can (or design equivalent) as a directional antenna to extend wifi network ranges. Sure, someone "invented" it first, but that doesn't mean it should be patented. Some "inventions" are trivial.
Well, yes, but that one has prior art due to the old Quaker oatmeal can wrapped with tin foil.
Prior art and what is "obvious" are fascinating aspects of patent law.