Partition-based pattern recognition system

1. An apparatus, comprising

a memory operative to maintain a knowledge element array comprising a plurality of knowledge elements; and

control logic operative to

partition the knowledge element array into one or more knowledge maps, wherein each knowledge map is associated with a partition identifier, and wherein each knowledge map comprises one or more knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier;

receive an input vector and a partition identifier; and

apply the input vector to the knowledge map corresponding to the partition identifier for recognition or learning operations,

wherein the control logic is further operative to receive a partition configuration for a selected partition, wherein the partition configuration comprises one or more attributes defining a distance calculation algorithm for computing distances between input vectors to the knowledge elements associated with a selected partition.

2. The apparatus of claim 1 wherein the control logic is further operative to

match the input vector to one or more knowledge elements of the knowledge map corresponding to the identified partition; and

return, responsive to one or more detected matches, one or more knowledge element attributes associated with corresponding knowledge elements.

3. The apparatus of claim 1 wherein the control logic is further operative to compare the input vector against one or more knowledge elements in one or more of the knowledge maps to conditionally match the input data vector to a category identifier.

4. The apparatus of claim 1 wherein the partition configuration further comprises one or more attributes defining the data type of the one or more operands, the bit-length of the one or more operands or the number of operands; and, one or more attributes defining a comparison technique for matching input vectors to the knowledge elements associated with a selected partition.

5. The apparatus of claim 4 wherein the control logic is further operative to match the input vector to one or more knowledge elements of the knowledge map corresponding to the identified partition using the distance calculation algorithm and comparison technique associated with the partition identifier.

6. The apparatus of claim 1 wherein the control logic is further operative to receive a partition configuration for a selected partition, wherein the partition configuration comprises one or more attributes defining the data type of the one or more operands, the bit-length of the one or more operands or the number of operands.

7. The apparatus of claim 1 wherein the control logic is further operative to receive a partition configuration for a selected partition, wherein the partition configuration comprises one or more attributes defining a comparison technique for matching input vectors to the knowledge elements associated with a selected partition.

8. The apparatus of claim 1 further comprising an application operative to select partitions and pass commands to the control logic identifying input vectors and selected partitions.

9. The apparatus of claim 1 further comprising an application operative to arrange partitions in a hierarchical relationship for a pattern recognition application.

10. The apparatus of claim 1 wherein one or more partitions are configured as elements of decisional nodes of a decision logic structure, wherein the decisional nodes comprise computer readable instructions that make one or more application programming interface calls that identify one or more partitions; and wherein the apparatus further comprise an application operative to traverse the decision logic structure, and use the output of a first partition as one or more operands of an input vector of a second partition operative to implement a higher order recognition operation.

11. The apparatus of claim 1 further comprising an application operative to access a first partition of the knowledge element array, and a second application operative to access a second partition of the knowledge element array.

12. The apparatus of claim 1 further comprising a second auxiliary memory;

and wherein the control logic is further operable to read knowledge elements from the second auxiliary memory into the memory, and read knowledge elements from the memory to the second auxiliary memory, as required in response to command identifying partitions of the knowledge element array.

13. The apparatus of claim 1 wherein the control logic is further operable to back up and restore knowledge elements from a persistent data store on a per-partition basis.

14. The apparatus of claim 1 wherein one or more knowledge elements further comprises user data; and wherein the control logic is operative to return, responsive to a match, the user data associated with one or more matching knowledge elements.

15. The apparatus of claim 2 wherein the control logic is further operable to maintain a counter for one or more knowledge elements, wherein the counters indicate the number of times a corresponding knowledge element matched to a received input vector.

16. The apparatus of claim 2 wherein the control logic is further operable to return corresponding identifiers for matching knowledge elements.

17. The apparatus of claim 1 wherein the control logic is further operable to delete one or more knowledge elements in response to a command.

18. The apparatus of claim 1 wherein the control logic is further operable to dynamically load logic defining a distance calculation algorithm associated with a partition.

19. The apparatus of claim 1 wherein the control logic is further operable to dynamically load logic defining a comparison technique algorithm associated with a partition.

20. A method comprising

maintaining a knowledge element array comprising a plurality of knowledge elements;

partitioning the knowledge element array into one or more knowledge maps, wherein each knowledge map is associated with a partition identifier, and wherein each knowledge map comprises one or more knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier;

receiving a partition configuration for a selected partition, wherein the partition configuration comprises one or more attributes defining a distance calculation algorithm for computing distances between input vectors to the knowledge elements associated with a selected partition;

receiving an input vector and a partition identifier, wherein the partition identifier corresponds to the selected partition; and

applying the input vector to the knowledge map corresponding to the partition identifier for recognition or learning operations using the distance calculation algorithm associated with the partition configuration of the selected partition.