Abstract: Image features are generated by performing wavelet transformations at sample points on images stored in electronic form. Multiple wavelet transformations at a point are combined to form an image feature vector. A prototypical set of feature vectors, or atoms, is derived from the set of feature vectors to form an “atomic vocabulary.” The prototypical feature vectors are derived using a vector quantization method (e.g., using neural network self-organization techniques) in which a vector quantization network is also generated. The atomic vocabulary is used to define new images. Meaning is established between atoms in the atomic vocabulary. High-dimensional context vectors are assigned to each atom. The context vectors are then trained as a function of the proximity and co-occurrence of each atom to other atoms in the image. After training, the context vectors associated with the atoms that comprise an image are combined to form a summary vector for the image.

Abstract: A method and system for intelligent control of external devices using a mammalian brain-like structure having three parts. The method and system include a computer-implemented neural network system which is an extension of the model-based adaptive critic design and is applicable to real-time control (e.g., robotic control) and real-time distributed control. Additional uses include data visualization, data mining, and other tasks requiring complex analysis of inter-relationships between data.

Abstract: In a data storage system, a cache is managed by a predictive cache management engine that evaluates cache contents and purges entries unlikely to receive sufficient future cache hits. The engine includes a single output back propagation neural network that is trained in response to various event triggers. Accesses to stored datasets are logged in a data access log; conversely, log entries are removed according to a predefined expiration criteria. In response to access of a cached dataset or expiration of its log entry, the cache management engine prepares training data. This is achieved by determining characteristics of the dataset at various past times between the time of the access/expiration and a time of last access, and providing these characteristics and the times of access as input to train the neural network. As another part of training, the cache management engine provides the neural network with output representing the expiration or access of the dataset.

Abstract: Method and apparatus for computer-aided assessment of organizational process or system. Method and apparatus are adapted to display questions on a computer to an assessor, who then inputs numerical and/or textual inputs relative to the assessor's perception of the process or system. Filtering techniques inhibit entry of unsupported numerical and/or textual inputs which may be untrue and/or exaggerated. Questions used in combination with additional questions provide more accurate assessment of the system or process, where the question combinations are already existing in a database. Evaluations of collected information are performed and additional corrective action questions are asked to identify degrees of risk and areas of weakness of the process or system.

Abstract: The invention is a set of processes for improving the interface of a self-organizing database. This type of database stores both questions and answers. The improved interface processes enable a user who has found a (first) question in the database to enter a new (second) question and label that second question as "more specific" than the first. When a user presses an "enter more specific question" command, the system enables the user to enter a question and then stores the question and labels it as more specific than the first, and links the two questions in the database. Then, when any user has found the first question, the system enables him to "go to" the linked, more specific question by entering a "go to more specific question" command. Further, when a user has found a question, the system can present a list of all the more specific questions linked to that question. The system also enables a user to find a first question and then enter a new (second) question and label it as a synonym of the first.

Abstract: A machine learning paradigm called Graph Transformer Networks extends the applicability of gradient-based learning algorithms to systems composed of modules that take graphs as inputs and produce graphs as output. Training is performed by computing gradients of a global objective function with respect to all the parameters in the system using a kind of back-propagation procedure. A complete check reading system based on these concept is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provides record accuracy on business and personal checks.

Abstract: A system is disclosed that provides a goal based learning system utilizing a rule based expert training system to provide a cognitive educational experience. The system provides the user with a simulated environment that presents a business opportunity to understand and solve optimally. Mistakes are noted and remedial educational material presented dynamically to build the necessary skills that a user requires for success in the business endeavor. The system utilizes an artificial intelligence engine driving individualized and dynamic feedback with synchronized video and graphics used to simulate real-world environment and interactions. Multiple "correct" answers are integrated into the learning system to allow individualized learning experiences in which navigation through the system is at a pace controlled by the learner.

Abstract: A method, apparatus and computer program product are provided for data mining including a feedback mechanism for monitoring performance of mining tasks. A user selected mining technique type is received for the data mining operation. A quality measure type is identified for the user selected mining technique type. The user selected mining technique type for the data mining operation is processed and a quality indicator is measured using the quality measure type. The measured quality indicator is displayed while processing the user selected mining technique type for the data mining operation.

Abstract: A computer operated apparatus estimates values needed by an optimizer in a database management system (DBMS). The DBMS has one or more tables for storing data, each table having zero or more columns of user-definable data types and zero or more associated user-defined routines (UDRs). The apparatus has a feature vector extractor connected to the database tables for converting the UDR inputs into a base type representation. A neural network receives the feature vector and generates estimated values which are provided to the optimizer of the DBMS. The neural network can be trained periodically using randomly generated queries, or it can be trained dynamically by capturing data generated during a query. During operation, the optimizer dynamically invokes the neural network to generate estimates such as selectivity and cost per call for determining optimum query search sequence.

Abstract: An improved method and apparatus for computing the speed of a robot. Specifically, the method and apparatus relate to calculating the speed of a robot based on the rated maximum speed of a joint of the robot and one of the current position and joint information. There are further aspects of the invention to calculate the speed of a scalar robot, a scalar robot with two degrees of freedom within a horizontal plane, a scalar robot with two degrees of freedom within a horizontal plane plus a front end shaft, a cylindrical robot having two degrees of freedom within a horizontal plane, and other configurations.

Abstract: Method and apparatus for computer-aided assessment of organizational process or system. Method and apparatus are adapted to display computer displayed questions to an assessor, who then inputs numerical inputs relative to the assessor's perception of the process or system. Filtering techniques inhibit entry of unsupported numerical inputs which may be untrue and/or exaggerated. Sequential questions used in combination provide more accurate assessment of the system or process, thereby enabling focused audits and/or inspections.

Abstract: A Multiple-Agent Hybrid Control Architecture (MAHCA) uses agents to analyze design, and implement intelligent control of distributed processes. A single agent can be configured to control a local process and a network of agents can be configured to control more complex distributed processes. Multiple agents interact through messages and can be either permanent or temporary. The network of agents interact to create an emergent global behavior. The network of agents support construction of closed-loop, autonomous systems which react to comply with two degrees of freedom: (1) a logic degree of freedom (useful for implementing switching between modes of control) and (2) an evolution degree of freedom (useful for implementing particular modes of control). Global behavior is emergent from individual agent behaviors and is achieved without central control through the imposition of global constraints on the network of individual agent behaviors (the logic agent network).

Abstract: A user initially judges whether each of pieces of information input as learning information is necessary or unnecessary, matrix elements of an affirmative metric signal indicating the records of the necessary information and matrix elements of a negative metric signal indicating the records of unnecessary information are calculated in a learning unit from a plurality of keywords attached to the necessary information and the unnecessary information. Thereafter, a plurality of keywords attached to each piece of information data input to be estimated are converted into a vector in a vector generating unit, and an affirmative score signal and a negative score signal are calculated from the vector and the affirmative and negative metric signals in a score calculating unit.

Abstract: Diagnostic systems utilizing a bayesian network model having link weights updated experientially include an algorithm for easily quantifying the strength of links in a Bayesian network, a method for reducing the amount of data needed to automatically update the probability matrices of the network on the basis of experiential knowledge, and methods and algorithms for automatically collecting knowledge from experience and automatically updating the Bayesian network with the collected knowledge. A practical exemplary embodiment provides a trouble ticket fault management system for a communications network. The exemplary embodiment is particularly appropriate for utilizing the automatic learning capabilities of the invention. In the exemplary embodiment, a communications network is represented as a Bayesian network where devices and communication links are represented as nodes in the Bayesian network.

Abstract: A task-based classification and analysis system includes an analysis software module and a user interface. The analysis software module establishes and maintains relationships between a plurality of databases or, in a preferred embodiment, hierarchical task lists. The user interface provides user inputs to the analysis software module such as budget information which is associated with particular elements of the databases. In consideration of historical data models, the user inputs and predetermined relationships between elements of the databases, the preferred system generates information products such as task-based budgets. Another preferred system coordinates task relationships between a plurality of software modules, such as a billing software module and a time entry software module.

Abstract: A system is disclosed that provides a goal based learning system utilizing a rule based expert training system to provide a cognitive educational experience. The system provides the user with a simulated environment that presents a business opportunity to understand and solve optimally. Mistakes are noted and remedial educational material presented dynamically to build the necessary skills that a user requires for success in the business endeavor. The system utilizes an artificial intelligence engine driving individualized and dynamic feedback with synchronized video and graphics used to simulate real-world environment and interactions. Multiple "correct" answers are integrated into the learning system to allow individualized learning experiences in which navigation through the system is at a pace controlled by the learner. A robust business model provides support for realistic activities and provides examples with process sensitive help that guides a student to intuitively develop the appropriate next step.

Abstract: A knowledge catalog includes a plurality of independent and parallel static ontologies to accurately represent a broad coverage of concepts that define knowledge. The actual configuration, structure and orientation of a particular static ontology is dependent upon the subject matter or field of the ontology in that each ontology contains a different point of view. The static ontologies store all senses for each word and concept. A knowledge classification system, that includes the knowledge catalog, is also disclosed. A knowledge catalog processor accesses the knowledge catalog to classify input terminology based on the knowledge concepts in the knowledge catalog. Furthermore, the knowledge catalog processor processes the input terminology prior to attachment in the knowledge catalog. The knowledge catalog further includes a dynamic level that includes dynamic hierarchies.

Abstract: A dynamic information architecture system and method for coupling one or more independent computer models to a computer expert system. A context manager establishes the couplings at run-time based upon a user's input goals and constraints. The context manager links available, but not readily compatible, computer models to solve for the user's input. An interface is provided for enabling the linkages through a structure of objects and object classes. Entity, aspect and process objects are created according to each computer model to provide a common interface. Large or small scale computer models can be accessed collectively, and complex systems can be thus be modeled. An event queue is automatically assembled to schedule events that form an event driven computer simulation.

Abstract: A multi-dimensional Gauss distribution generating portion obtains one vector from a multi-dimensional Gauss probability distribution and outputs it as an initial momentum. An approximated energy function differential generating portion generates a gradient of an approximated energy function. A Hamilton equation integral portion integrates a Hamilton equation using a gradient of an approximated energy function for a predetermined integral period with initial conditions of an initial state and an initial momentum. The result of the integration as the state of the next alternative is output from a next alternative generating apparatus. The state is used with a high probability corresponding to the difference of approximated energy calculated by an approximated energy difference calculating portion and the difference between a value of an original energy in the present state and that in the next state.

Abstract: A method and system for natural language processing including using a trained neural network having a plurality of baseline nodes. A connection weight between any selected pair of baseline nodes is described by text strings within the selected pair of nodes. A plurality of text messages are received from a preselected source. For each received message, a non-baseline node associated with a selected one of the baseline nodes is created. A connection weight between any non-baseline node and the associated baseline node is described by the text string within the baseline node and the received text message.