In former articles of this blog, it has been seen that people usually invent or solve problems :
* thanks to their intuition, their common sense (see the article named "Wine tradeshow") ;
* thanks to the deep knowledge of their environment (see for instance the black hornets' problem) ;
* thanks to the deep knowledge of their specialty field, to tests and experiments (as usually engineers and technicians do in the industry), computerized simulations included.
I stress that TRIZ practitionners invent and solve problems. Some other systematic methods can help, too.
Let us try to envision the future : will it be possible one day that machines invent like human beings ?
Actually these days have already come : since 1995 John Koza's computers have produced several inventions (i.e. patentable designs), mainly in the field of electrical circuits, but not only.
Thus John Koza (see picture above), a world specialist in his field, developed genetic programming for automated design, whereas before his pioneering discovery, people used genetic programming for global optimization. In other words, instead of changing only physical parameters of a physical model, Koza makes the design, the structure itself of an object or a system evolving through genetic algorithms. This genetical evolution mimics somehow the process of darwinian evolution of life. Based on this work, a colleague of his from NASA, used such algorithms to design a microwave antenna (see below the picture of an evolved antenna), which fulfilled all the technical requirements of its spatial mission (http://www.genetic-programming.com/c2003jasonlohn20031124talk.pdf). The shape of the resulting antenna was unexpected (http://www.genetic-programming.com/coursemainpage.html) , and it is believed that no human being could have thought of this strange structure.
* thanks to their intuition, their common sense (see the article named "Wine tradeshow") ;
* thanks to the deep knowledge of their environment (see for instance the black hornets' problem) ;
* thanks to the deep knowledge of their specialty field, to tests and experiments (as usually engineers and technicians do in the industry), computerized simulations included.
I stress that TRIZ practitionners invent and solve problems. Some other systematic methods can help, too.
Let us try to envision the future : will it be possible one day that machines invent like human beings ?
Actually these days have already come : since 1995 John Koza's computers have produced several inventions (i.e. patentable designs), mainly in the field of electrical circuits, but not only.
Thus John Koza (see picture above), a world specialist in his field, developed genetic programming for automated design, whereas before his pioneering discovery, people used genetic programming for global optimization. In other words, instead of changing only physical parameters of a physical model, Koza makes the design, the structure itself of an object or a system evolving through genetic algorithms. This genetical evolution mimics somehow the process of darwinian evolution of life. Based on this work, a colleague of his from NASA, used such algorithms to design a microwave antenna (see below the picture of an evolved antenna), which fulfilled all the technical requirements of its spatial mission (http://www.genetic-programming.com/c2003jasonlohn20031124talk.pdf). The shape of the resulting antenna was unexpected (http://www.genetic-programming.com/coursemainpage.html) , and it is believed that no human being could have thought of this strange structure.
No comments:
Post a Comment