Data processing: artificial intelligence – Machine learning – Genetic algorithm and genetic programming system
Reexamination Certificate
1999-09-10
2003-01-07
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Machine learning
Genetic algorithm and genetic programming system
Reexamination Certificate
active
06505180
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Technical Field
This invention generally relates to genetic algorithms, information encoding and information encryption. Specifically, this invention relates to the structure of genetic algorithms and how that structure can be made to more accurately parallel the natural genetic paradigm; the structure of complex data set encoding, in particular, digitally volume rendered data sets; and absolute secure information encryption methods. The invention is particularly applicable to the range of problems wherein a more natural genetic-like implementation of the conventional genetic algorithm, the increased efficiency of information encoding in the form of synthetic genes and/or encryption of information utilizing synthetic genes may be found as a solution to these problems.
2. Background
Genetic algorithms have been used for decades for solving problems in many fields of research, development and implementation, most notably, the sciences and engineering. A central appeal of genetic algorithms has been the notion of novel solutions to otherwise intractable problems. For many of the same reasons that great fascination and often amazement exists for natural biological genetics, researchers have often found unique and unexpected solutions to various problem sets through the use of genetic algorithms.
In large part, genetic algorithms possess their power and effectiveness precisely because they mimic a process that, by the reckoning of current science, has been replaying itself successfully for billions of years.
Therefor, it would be desirable to provide a higher degree of transcription of natural genetic principles and apply this to the field of genetic algorithms.
As technological advances create increased opportunities for data utilization and visualization, so too are there increased strains on the prior art of encoding and retrieving such information. In particular, the field of three dimensionally viewed objects and their manipulation in digital environments have become increasingly important in medicine, education and science. Virtual reality is fast becoming an accepted, and in some cases the only, means of achieving effective communication, education and simulation. Unfortunately, information current methods of information encoding are largely inefficient and cumbersome.
Therefor, it would be desirable to provide more efficient and compact method of information encoding.
Because of exponentially increasing digital traffic, both on and off the internet, the need for totally secure forms of information exchange has become paramount. With the advent of ever more powerful computers available to almost everyone, computing intensive algorithms for data encryption, as well as the decoding of such encryption, have now become practically feasible.
The rapid rise of technical tools, more common place expertise, and the sheer power of readily available computers have spawned a new class of users whose motives are not always honorable. “Code cracking” and computer “hacking” in general have become new avenues for entertainment, achievement and criminal activity. To date, very few, if any, absolutely secure information encyrption algorithms exist.
Genetic algorithms are generally parallel processes that develop populations of individual data objects, usually developing binary character strings into new populations of the same data type, using methods that mimic biological genetics, such as recombination, or crossover, and a proportional reproductive schema based on the notion of Darwin's survival of the fittest. Such algorithms start with some initial population of data objects, usually created in some pseudo random fashion. These data objects are then evaluated iteratively for fitness as it pertains to the problem at hand and genetic like operations are performed on various data objects within the population to evolve a new population.
John Holland of the University of Michigan developed the initial concept of genetic algorithms for fixed-length binary character strings in Adaptation in Artificial and Natural Systems, by Professor John H. Holland, 1975. Subsequent and significant works in genetic algorithms and genetic classifier systems may be referenced in Grefenstette, Proceedings of the Second International Conference on Genetic Algorithms, 1987; M. Srinivas, et al., “Genetic Algorithms: A Survey”, Computer, vol. 27, No. 6, pages 17-26, Jun. 1994; Goldberg, Genetic Algorithms, pages 10-20, 80-139, Addison Wesley 1989; W. B. Dress, “Darwinian Optimization of Synthetic Neural Systems”, IEEE First International Conference on Neural Networks, San Diego, Jun. 1987, vol. No. 3, pages 769-775; Schaffer et al., An Adaptive Crossover Distribution Mechanism for Genetic Algorithms, Proceedings of the 2
nd
International Conference on Genetic Algorithms, Jul. 28-31, 1987, pages 36-40; and Melanie Mitchell, “An Introduction to Genetic Algorithms”, pages 87-113, MIT Press 1996.
Several improvements have been made to Holland's basic premise over the ensuing years, but none has addressed the lack of parallelism between these genetic algorithms and their natural genetic paradigm, namely that in conventional genetic algorithms, sexual-like recombination or crossover, regardless of its geometry or relative sophistication, occurs on a population member or data object only in its final form, that is its phenotypical form. This is the biological equivalent, for example, of grafting the legs of a very fast runner onto the body of a person having great upper body strength in order to achieve some environmental fitness objective.
Indeed, all efforts toward improvements in genetic algorithms have been made on the same plane. That is to say, the prior art procedures of crossover and fitness selection in genetic algorithms, regardless of their variations, are performed on the same level of member development, without regard for the complexities surrounding the integrated behavior of the genomic regulatory systems underlying ontogeny, or in other words, the unfolding of events involved in an organism changing gradually from a simple to a more complex level.
In reality, biological sexual recombination occurs at the genotypical level, where the genotype is a group of organisms sharing a specific genetic constitution. The phenotype, the constitution of an organism as determined by the interaction of its genetic constitution and the environment, is the phase in the overall scheme of things where natural selection occurs.
In fact, the two procedures, biological crossover or sexual recombination and fitness selection, occur not only at very different levels but on an incredibly different scale.
“It has been estimated that the sperm that fertilized all the eggs from which the present human population of the world developed could be packed into a container smaller than an eraser on a pencil . . . the biologically inherited qualities of human beings—the similarities as well as the differences that distinguish one human from another and from all other living things—have their basis in a minute mass of sperm . . . ”. Adrian M. Srb, “General Genetics”, pages 2-26, 265-284., W. H. Freeman and Company 1965.
Volume rendering is a computer graphics technique whereby the object or phenomenon of interest is sampled or subdivided into many cubic building blocks, called voxels, or volume elements. A voxel is the three dimensional (hereafter 3-D) counterpart of the two dimensional (hereafter 2-D) pixel and is a measure of unit volume. Each voxel carries one or more values for some measured or calculated property of the volume and is typically represented by a unit cube. The 3-D voxel sets are assembled from multiple 2-D images, and are displayed by projecting these images into 2-D pixel space where they are stored in a frame buffer. Volumes rendered in this manner have been likened to a translucent suspension of particles in 3-D space.
In surface rendering, the volumetric data must first be converted into geometric primitives, by a process such as isosurfacing, isocontouring, surface extractio
Chapman and Cutler
Davis George B.
Wm. L. Crowley & Associates, Inc.
LandOfFree
Information encoding and retrieval through synthetic genes does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Information encoding and retrieval through synthetic genes, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information encoding and retrieval through synthetic genes will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3070893