Back to results
US 8,014,591 B2Granted
Robust pattern recognition system and method using socratic agents
AssigneeAurilab, LLC
InventorJames K. Baker
Life of this patent
Estimating the term…
Loading patent…
Loading patent…
AssigneeAurilab, LLC
InventorJames K. Baker
Estimating the term…
A computer-implemented pattern recognition method, system and program product, the method comprising in one embodiment: creating electronically a linkage between a plurality of models within a classifier module within a pattern recognition system such that any one of said plurality of models may be selected as an active model in a recognition process; creating electronically a null hypothesis between at least one model of said plurality of linked models and at least a second model among said plurality of linked models; accumulating electronically evidence to accept or reject said null hypothesis until sufficient evidence is accumulated to reject said null hypothesis in favor of one of said plurality of linked models or until a stopping criterion is met; and transmitting at least a portion of the electronically accumulated evidence or a summary thereof to accept or reject said null hypothesis to a pattern classifier module.
A computer-implemented pattern recognition method, comprising:
Dependencies are inferred from the claim text.
From the US filing history. Foreign family members aren’t listed.
Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation: validation; Performance…
Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation: generating training…
Computing and data processing
About 3 years 8 months of term left · ends 6 Jul 2030 (estimated)
Although much progress has been made in pattern recognition systems, with the huge and rapidly growing amount of information that needs to be processed there is a need for further improvement, especially for systems that can handle a large quantity of data.
Finally, it is desirable in some embodiments to have a recognition system that is so robust that it can find and correct its own errors.
These objectives and others are addressed by the present invention.
The invention will be described by reference to the following Figures and diagrams:
FIG. 1 is a block diagram of a higher level Socratic agent acquiring knowledge about and controlling an associated lower-level classifier module.
FIG. 2 is a block diagram of a classifier module and a plurality of associated Socratic agents.
Explanation of Special Terms:
A knowledge source is an object that includes a mechanism for knowledge representation and either a mechanism for knowledge acquisition or a collection of built-in base knowledge or both.
A classifier module is a processing module that receives samples of data features as input and generates classification results as output. The classification results may, for example, comprises an indication of the class best matching the input data or may comprises a score for each class indicating how well the data matches the given class. Implicitly, any pattern classifier has a representation of knowledge about the class it identifies. Generally, a classifier module is trainable and hence has a mechanism for knowledge acquisition. Therefore, any pattern classifier is a kind of knowledge source. It is to be understood that a classifier module is not necessarily implemented as a separate software module or as a separate piece of hardware. A single software module may implement a plurality of classifier modules. A single software module may also implement one or more lower-level modules as well as a higher-level module. In particular, a Socratic controller, which comprises a higher-level classifier module associated with a plurality of lower-level classifier modules may be implemented as either a single software module or as a plurality of software modules.