Empirical data modeling
Methods, apparatuses and systems directed to pattern identification and pattern recognition. In some particular implementations, the invention provides a flexible pattern recognition platform including pattern recognition engines that can be dynamically adjusted to implement specific pattern recognition configurations for individual pattern recognition applications. In some implementations, the present invention also provides for a partition configuration where knowledge elements can be grouped and pattern recognition operations can be individually configured and arranged to allow for multi-level pattern recognition schemes.
1. A pattern recognition system, comprising memory and one or more processors operatively coupled to the memory and configured in a learning-enabled mode to:
receive a plurality of input vectors;
generate a knowledge element for each of first ones of the input vectors in unsupervised operations, each knowledge element including a point in an n-dimensional space represented by the corresponding input vector and having an influence field associated therewith, wherein the influence fields of the respective knowledge elements are substantially equal, and wherein the knowledge elements represent a single category of data;
match each of second ones of the input vectors to at least one of the knowledge elements within the influence field of which the second input vector lies; and
maintain a count for each knowledge element representing a number of the input vectors corresponding to that knowledge element, the counts collectively representing statistical regularities of patterns represented by the input vectors;
wherein the one or more processors are further configured to, in a learning-disabled mode in which generation of further knowledge elements is not allowed, match third ones of the input vectors to the knowledge elements.
2. The pattern recognition system of claim 1 wherein the plurality of knowledge elements forms a first one of a plurality of partitions of a knowledge element array maintained in the memory, and wherein the one or more processors are further configured to receive a partition identifier and a partition configuration corresponding to the first partition, the partition configuration identifying a comparison technique for comparing the second input vectors to the knowledge elements of the partition.
3. The pattern recognition system of claim 2 wherein the one or more processors are further configured to combine results from the plurality of partitions to achieve a higher level result by connecting operations of the plurality of partitions in a hierarchical relationship or a serial relationship.
4. The pattern recognition system of claim 1 wherein the one or more processors are further configured to adjust the influence fields of the knowledge elements substantially equally as new knowledge elements are generated, thereby constraining a total number of knowledge elements.
5. The pattern recognition system of claim 1 wherein the one or more processors are further configured to limit generation of knowledge elements to a maximum number.
6. The pattern recognition system of claim 1 wherein the one or more processors are further configured to delete particular ones of the knowledge elements based on the corresponding counts.
7. A pattern recognition system, comprising memory and one or more processors operatively coupled to the memory and configured to:
receive a plurality of input vectors;
process the input vectors by masking or weighting one or more components of at least some of the input vectors;
generate a knowledge element for each of first ones of the input vectors, each knowledge element including a point in an n-dimensional space represented by the corresponding input vector and having an influence field associated therewith, wherein the influence fields of the respective knowledge elements are substantially equal, wherein the knowledge elements represent a single category of data, and wherein the one or more processors are configured to limit generation of knowledge elements to a maximum number;
match each of second ones of the input vectors to at least one of the knowledge elements within the influence field of which the second input vector lies; and
maintain a count for each knowledge element representing a number of the input vectors corresponding to that knowledge element, the counts collectively representing statistical regularities of patterns represented by the input vectors.
8. The pattern recognition system of claim 7 wherein the one or more processors are configured to generate the knowledge elements in unsupervised operations.
9. The pattern recognition system of claim 7 wherein the first and second input vectors comprise training vectors, and the one or more processors are configured to generate the knowledge elements in supervised operations.
10. The pattern recognition system of claim 7 wherein the one or more processors are configured to generate the knowledge elements and associated counts in a learning-enabled mode, and to match third ones of the input vectors to the knowledge elements in a learning-disabled mode in which generation of further knowledge elements is not allowed.
11. The pattern recognition system of claim 7 wherein the plurality of knowledge elements forms a first one of a plurality of partitions of a knowledge element array maintained in the memory, and wherein the one or more processors are further configured to receive a partition identifier and a partition configuration corresponding to the first partition, the partition configuration identifying a comparison technique for comparing the second input vectors to the knowledge elements of the partition.
12. The pattern recognition system of claim 11 wherein the one or more processors are further configured to combine results from the plurality of partitions to achieve a higher level result by connecting operations of the plurality of partitions in a hierarchical relationship or a serial relationship.
13. The pattern recognition system of claim 7 wherein the one or more processors are further configured to adjust the influence fields of the knowledge elements substantially equally as new knowledge elements are generated, thereby constraining a total number of knowledge elements.
14. The pattern recognition system of claim 7 wherein the one or more processors are further configured to delete particular ones of the knowledge elements based on the corresponding counts.
15. A pattern recognition system, comprising memory and one or more processors operatively coupled to the memory and configured to:
generate a plurality of input vectors using previously digitized information or information generated by one or more sensors;
generate a knowledge element for each of first ones of the input vectors, each knowledge element including a point in an n-dimensional space represented by the corresponding input vector and having an influence field associated therewith, wherein the influence fields of the respective knowledge elements are substantially equal, and wherein the knowledge elements represent a single category of data;
match each of second ones of the input vectors to at least one of the knowledge elements within the influence field of which the second input vector lies;
maintain a count for each knowledge element representing a number of the input vectors corresponding to that knowledge element, the counts collectively representing statistical regularities of patterns represented by the input vectors; and
delete particular ones of the knowledge elements based on the corresponding counts.
16. The pattern recognition system of claim 15 wherein the one or more processors are configured to generate the knowledge elements in unsupervised operations.
17. The pattern recognition system of claim 15 wherein the first and second input vectors comprise training vectors, and the one or more processors are configured to generate the knowledge elements in supervised operations.
18. The pattern recognition system of claim 15 wherein the one or more processors are configured to generate the knowledge elements and associated counts in a learning-enabled mode, and to match third ones of the input vectors to the knowledge elements in a learning-disabled mode in which generation of further knowledge elements is not allowed.
19. The pattern recognition system of claim 15 wherein the plurality of knowledge elements forms a first one of a plurality of partitions of a knowledge element array maintained in the memory, and wherein the one or more processors are further configured to receive a partition identifier and a partition configuration corresponding to the first partition, the partition configuration identifying a comparison technique for comparing the second input vectors to the knowledge elements of the partition.
20. The pattern recognition system of claim 19 wherein the one or more processors are further configured to combine results from the plurality of partitions to achieve a higher level result by connecting operations of the plurality of partitions in a hierarchical relationship or a serial relationship.
21. The pattern recognition system of claim 15 wherein the one or more processors are further configured to adjust the influence fields of the knowledge elements substantially equally as new knowledge elements are generated, thereby constraining a total number of knowledge elements.
22. The pattern recognition system of claim 15 wherein the one or more processors are further configured to limit generation of knowledge elements to a maximum number.
23. The pattern recognition system of claim 15 wherein the one or more processors are further configured to process the input vectors by masking or weighting one or more components of at least some of the input vectors.