Other

Telcordia Technologies, Inc.

Tt Government Solutions, Inc.

Clifford Behrens

Devasis Bassu

Dennis Egan

Hyong-Sop Shim

Hanan Luss

George W Seymour

Martin Eiger

Hiralal Agrawal

Balakrishnan Dasarathy

Tamra Carpenter

Paul Seymour

Mark Rosenstein

Carol Lochbaum

Yu-Yun Ho

A computer-enabled method of learning the behavior of a program. A processor can execute a target program during a learning interval while varying a plurality of stimuli provided to the target program so as to produce a multiplicity of different sequences of events which differ in combinations of types of events in respective sequences, orders in which the types of events occur in respective sequences, or in the combinations and in the orders in which the types of events occur. The multiplicity of event sequences can be recorded, and a second program can be executed by a processor to: determine a plurality of clusters based on similarities between the event sequences in their entirety; and determine a plurality of signatures corresponding to the plurality of clusters. Each signature can be the longest common subsequence of all sequences in the respective cluster and thus representative of the cluster. In such method, each of the plurality of signatures can be a benchmark representative of acceptable behavior of the target program.

Learning program behavior for anomaly detection

A consensus-based knowledge validation and analysis system provides a way to increase use of collaboration tools among panels of experts by providing a system for analyzing and validating the responses of such experts to a set of questions. The system uses a set of response data input by a panel of experts with respect to a particular subject matter formatted in accordance with a data model as input. The response data set is used to estimate an empirical point estimate matrix indicative of the amount of agreement in the responses on all items between the panelists. The empirical point estimate matrix is used to estimate the saliency of the subject matter to panelists, the competency of each panelist and a consensus model of correct answers is based on the estimated competency of each panelist and the of responses for each item in the response data set. This consensus model is used to generate a knowledge map to aid visualization of the consensus data and encourage further collaboration and consensus building. The method is implemented in a web-based system that enables users of collaboration tools to send response data sets to the tool via the Internet or virtual private network and to likewise retrieve knowledge maps, panelist information and consensus data. An interactive feature enables users/panelists to collaborate with other panelists using the knowledge map as an interface to one or more collaboration tools such as instant messaging.

System and method for consensus-based knowledge validation, analysis and collaboration

Indexing, searching, and retrieving the content of speech documents (including but not limited to recorded books, audio broadcasts, recorded conversations) is accomplished by finding and retrieving speech documents that are related to a query term at a conceptual level, even if the speech documents does not contain the spoken (or textual) query terms. Concept-based cross-media information retrieval is used. A term-phoneme/document matrix is constructed from a training set of documents. Documents are then added to the matrix constructed from the training data. Singular Value Decomposition is used to compute a vector space from the term-phoneme/document matrix. The result is a lower-dimensional numerical space where term-phoneme and document vectors are related conceptually as nearest neighbors. A query engine computes a cosine value between the query vector and all other vectors in the space and returns a list of those term-phonemes and/or documents with the highest cosine value.

Concept based cross media indexing and retrieval of speech documents

The use of latent semantic indexing (LSI) for information retrieval and text mining operations is adapted to work on large heterogeneous data sets by first partitioning the data set into a number of smaller partitions having similar concept domains. A similarity graph network is generated in order to expose links between concept domains which are then exploited in determing which domains to query as well as in expanding the query vector. LSI is performed on those partitioned data sets most likely to contain information related to the user query or text mining operation. In this manner LSI can be applied to datasets that heretofore presented scalability problems. Additionally, the computation of the singular value decomposition of the term-by-document matrix can be accomplished at various distributed computers increasing the robustness of the retrieval and text mining system while decreasing search times.

Information retrieval and text mining using distributed latent semantic indexing

A method for placing equipment in a network begins with a baseline network on which demographic data is analyzed statistically to produce a demographically-driven driven demand forecast model. The demand forecast model yields a predicted demand for each census block group. This can be visualized as one map layer with a GIS. The layout method takes the predicted demands as input and produces new attribute information for network nodes, the central office and cross-connects. By the layout method the cost of placing equipment at available placement sites or nodes is minimized based on various constraints, e.g., the capacity of the equipment to be placed, the distance a user is located from a potential node. Once the minimum cost for a given set of placement sites, subscribers, equipment capacity, and other known constraints is determined the equipment are then placed at these sites. These include the number and type of DSLAMs assigned to each node. These can be visualized with the GIS by clicking-on a node to expose the equipment deployed there.

Digital subscriber line network deployment method

Techniques for using latent semantic structure of textual content ascribed to the items to provide automatic recommendations to the user. A user inputs a selected item and, in turn, a latent semantic algorithm is applied to the user selection and the textual content of the items in a database to generate a conceptual similarity between the selection and the items. A set of nearest items to the selected item is provided as a recommendation to the user of other items that may be of particular interest or relevance to the user's original selection based upon the conceptual similarity measure.

Growing community of inventors

Madison, NJ, United States of America

Clifford Behrens