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 multilevel pattern recognition schemes.
1. A pattern recognition system comprising
a memory;
one or more processors;
logic operable to cause the one or more processors to:
receive, in a learning process, a plurality of input vectors;
iteratively process the input vectors to compute a knowledge map, wherein the processing comprises
determining whether an input vector falls within an influence field of an existing knowledge element of the knowledge map;
if so, initially rejecting the input vector;
if not, entering the input vector into the knowledge map as a knowledge element and conditionally adjusting one or more influence fields of other knowledge elements; and
further computing the knowledge map by processing one or more of the initially rejected input vectors.
2. A pattern recognition system as in claim 1 , further comprising logic operable to cause the one or more processors to:
optionally adjust the influence fields in the map through a half-learning process; and
use the knowledge map created in the learning process to determine whether other input vectors match those in the knowledge map.
3. A pattern recognition system as in claim 1 , further comprising logic operable to cause the one or more processors to:
use the knowledge map created in the learning process to determine whether other input vectors match those in the knowledge map in a hierarchical manner.
4. A method comprising
receiving, by a pattern recognition device in a learning process, a plurality of input vectors;
iteratively processing, by the pattern recognition device, the input vectors to compute a knowledge map, wherein the processing comprises
determining whether an input vector falls within an influence field of an existing knowledge element of the knowledge map;
if so, initially rejecting the input vector;
if not, entering the input vector into the knowledge map as a knowledge element and conditionally adjusting one or more influence fields of other knowledge elements; and
further computing, by the pattern recognition device, the knowledge map by processing one or more of the initially rejected input vectors.
5. The method of claim 4 , further comprising
receiving an input vector in a half-learning mode; and
optionally adjusting the influence fields in the knowledge map based on the received input vector without creating a knowledge element based on the input vector.
6. The method of claim 5 further comprising using the knowledge map created in the learning process to determine whether other input vectors match those in the knowledge map.
7. The method of claim 4 , further comprising using the knowledge map created in the learning process to determine whether other input vectors match those in the knowledge map.
8. An apparatus, comprising
a memory operative to maintain a knowledge element array comprising a plurality of knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier; and
control logic operative to
receive configuration information defining weights to be applied to one or more operands;
receive an input vector;
match the input vector to one or more knowledge elements of the knowledge map by computing the aggregate distance between the input vector and the one or more knowledge elements, wherein the aggregate distance is influenced by the weights associated with the one or more operands; and
return, responsive to one or more detected matches, one or more category identifiers associated with corresponding knowledge elements.
9. An apparatus, comprising
a memory operative to maintain a knowledge element array comprising a plurality of knowledge elements, each knowledge element comprising one or more operands collectively defining a data vector and a category identifier; and
control logic operative to
receive configuration information defining one or more operands to be masked;
receive an input vector;
match the input vector to one or more knowledge elements of the knowledge map by computing the aggregate distance between the input vector and the one or more knowledge elements, wherein the one or more masked operands are omitted from computing the aggregate distance; and
return, responsive to one or more detected matches, one or more category identifiers associated with corresponding knowledge elements.
10. Logic encoded in one or more tangible media for execution and when executed operable to:
receive, in a learning process, a plurality of input vectors;
iteratively process the input vectors to compute a knowledge map, wherein to process the input vectors, the logic is further operable to
determine whether an input vector falls within an influence field of an existing knowledge element of the knowledge map;
if so, initially reject the input vector;
if not, enter the input vector into the knowledge map as a knowledge element and conditionally adjust one or more influence fields of other knowledge elements; and
further compute the knowledge map by processing one or more of the initially rejected input vectors.