Enhanced learning and recognition operations for radial basis functions
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 a memory configured to maintain a knowledge element array comprising a plurality of knowledge elements; and logic configured to:
partition the knowledge element array into one or more knowledge maps, each knowledge map having an associated partition identifier and comprising one or more of the 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 particular partition identifier;
receive a partition configuration corresponding to the particular partition identifier and identifying a comparison technique for comparing the input vector and the data vectors of the one or more knowledge elements of a particular one of the one or more knowledge maps associated with the particular partition identifier; and
process the input vector by applying the input vector to the particular knowledge map for a recognition operation or a learning operation using the comparison technique.
2. The pattern recognition system of claim 1 wherein the input vector comprises image data representing an image, video data representing a video, or waveform data representing a waveform.
3. The pattern recognition system of claim 2 wherein the logic is further configured to identify an object in the image, to identify an object in the video, or to identify a waveform characteristic in the waveform.
4. The pattern recognition system of claim 1 wherein the logic is configured to process the input vector for a recognition operation by masking or weighting one or more operands of the input vector.
5. The pattern recognition system of claim 1 wherein the logic is configured to process the input vector for an unsupervised learning operation.
6. The pattern recognition system of claim 1 wherein the input vector is derived from previously digitized information or from information generated by one or more sensors.
7. The pattern recognition system of claim 1 further comprising feature extraction logic configured to generate the input vector using the previously digitized information or the information generated by one or more sensors.
8. The pattern recognition system of claim 1 wherein the logic is configured to process the input vector for a recognition operation and generate one or more of a spatial recognition result, a temporal recognition result, or a probabilistic recognition result.
9. A computer program product for recognizing patterns using a knowledge element array comprising a plurality of knowledge elements, the computer program product comprising one or more non-transitory computer-readable media having computer program instructions stored therein, the computer program instructions being configured, when executed, to cause one or more computing devices to:
partition the knowledge element array into one or more knowledge maps, each knowledge map having an associated partition identifier and comprising one or more of the 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 particular partition identifier;
receive a partition configuration corresponding to the particular partition identifier and identifying a comparison technique for comparing the input vector and the data vectors of the one or more knowledge elements of a particular one of the one or more knowledge maps associated with the particular partition identifier; and
process the input vector by applying the input vector to the particular knowledge map for a recognition operation or a learning operation using the comparison technique.
10. The computer program product of claim 9 wherein the input vector comprises image data representing an image, video data representing a video, or waveform data representing a waveform.
11. The computer program product of claim 10 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to identify an object in the image, to identify an object in the video, or to identify a waveform characteristic in the waveform.
12. The computer program product of claim 9 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to process the input vector for a recognition operation by masking or weighting one or more operands of the input vector.
13. The computer program product of claim 9 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to process the input vector for an unsupervised learning operation.
14. The computer program product of claim 9 wherein the input vector is derived from previously digitized information or from information generated by one or more sensors.
15. The computer program product of claim 9 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to generate the input vector using the previously digitized information or the information generated by one or more sensors.
16. The computer program product of claim 9 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to process the input vector for a recognition operation and generate one or more of a spatial recognition result, a temporal recognition result, or a probabilistic recognition result.
17. A pattern recognition system comprising a memory configured to maintain a knowledge element array comprising a plurality of knowledge elements; and logic configured to:
partition the knowledge element array into a plurality of knowledge maps, each knowledge map having an associated partition identifier and comprising one or more of the knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier;
receive a plurality of input vectors each having an associated one of the partition identifiers;
receive one or more partition configurations each corresponding to one or more of the partition identifiers, each partition configuration identifying a comparison technique for comparing the input vectors and the data vectors of the one or more knowledge elements of the knowledge maps associated with the one or more partition identifiers corresponding to the partition configuration;
process each input vector by applying the input vector to the knowledge map for the partition identifier associated with the input vector for a recognition operation using the comparison technique; and
combine results of a plurality of the recognition operations performed using different ones of the knowledge maps to achieve a higher level recognition result.
18. The pattern recognition system of claim 17 wherein the logic includes decisional logic configured to connect the recognition operations from the different knowledge maps in a hierarchical relationship.
19. The pattern recognition system of claim 17 wherein the logic includes decisional logic configured to connect the recognition operations from the different knowledge maps in a serial relationship.
20. The pattern recognition system of claim 17 wherein each of the knowledge maps is configured for processing input vectors corresponding to one of a plurality of different data types.
21. The pattern recognition system of claim 20 wherein the different data types are derived from a plurality of different sensor types.
22. The pattern recognition system of claim 20 wherein the plurality of different data types comprise any of image data, video data, audio data, waveform data, chemical data, text data, or binary data.
23. The pattern recognition system of claim 17 wherein at least some of the knowledge maps are configured for processing input vectors corresponding to the same data type.
24. The pattern recognition system of claim 17 wherein at least one of the results of the recognition operations comprises a result vector derived with reference to opaque user data associated with one or more knowledge elements, and wherein the logic is configured to combine the results of the recognition operations by applying the result vector to a subsequent one of the knowledge maps.
25. A computer program product for recognizing patterns using a knowledge element array comprising a plurality of knowledge elements, the computer program product comprising one or more non-transitory computer-readable media having computer program instructions stored therein, the computer program instructions being configured, when executed, to cause one or more computing devices to:
partition the knowledge element array into a plurality of knowledge maps, each knowledge map having an associated partition identifier and comprising one or more of the knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier;
receive a plurality of input vectors each having an associated one of the partition identifiers;
receive one or more partition configurations each corresponding to one or more of the partition identifiers, each partition configuration identifying a comparison technique for comparing the input vectors and the data vectors of the one or more knowledge elements of the knowledge maps associated with the one or more partition identifiers corresponding to the partition configuration;
process each input vector by applying the input vector to the knowledge map for the partition identifier associated with the input vector for a recognition operation using the comparison technique; and
combine results of a plurality of the recognition operations performed using different ones of the knowledge maps to achieve a higher level recognition result.
26. The computer program product of claim 25 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to connect the recognition operations from the different knowledge maps in a hierarchical relationship.
27. The computer program product of claim 25 wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to connect the recognition operations from the different knowledge maps in a serial relationship.
28. The computer program product of claim 25 wherein each of the knowledge maps is configured for processing input vectors corresponding to one of a plurality of different data types.
29. The computer program product of claim 28 wherein the different data types are derived from a plurality of different sensor types.
30. The computer program product of claim 28 wherein the plurality of different data types comprise any of image data, video data, audio data, waveform data, chemical data, text data, or binary data.
31. The computer program product of claim 25 wherein at least some of the knowledge maps are configured for processing input vectors corresponding to the same data type.
32. The computer program product of claim 25 wherein at least one of the results of the recognition operations comprises a result vector derived with reference to opaque user data associated with one or more knowledge elements, and wherein the computer program instructions are further configured, when executed, to cause the one or more computing devices to combine the results of the recognition operations by applying the result vector to a subsequent one of the knowledge maps.