Gm Global Technolgoy Operations, Inc.

Hrl Laboratories, LLC

Dnyanesh Rajpathak

John Anthony Cafeo

Soumen De

Joseph A Donndelinger

Prakash M Peranandam

Tsai-Ching Lu

Layne K Wiggins

Jiejun Xu

Chaitanya Sankavaram

Prakash Mohan Peranandam

Vineet Khare

Azeem Sarwar

Ramesh Sethu

Dean G Sorrell

Xiangxing Lu

Rahul G Chougule

Ravi S Sambangi

Susan H Owen

Kay L Darden

Xinli Wang

Baiyang Wang

A system processes unstructured data to identify a plurality of subsets of text in a set of text in the unstructured data and determines, for a subset from the plurality of subsets, probabilities based on a position of the subset in the set of text, a part of speech (POS) of each word in the subset, and POSs of one or more words on left and right hand sides of the subset, a number of the one or more words being selected based on a length of the set of text. The system generates a feature vector for the subset, the feature vector including the probabilities and additional features of the subset; and classifies, using a classifier, the subset into one of a plurality of classes based on the feature vector for the subset, the plurality of classes representing an ontology of a domain of knowledge.

Deep learning based automatic ontology extraction to detect new domain knowledge

Described is a system for identification of correlations in customer observables (COs). The system extracts key phrases representing COs from textual inputs from multiple data sources, wherein the COs are related to a consumer product. A unified hypergraph is constructed that models co-occurrences of COs. The unified hypergraph includes nodes and types of hyperedges connecting the nodes, where COs are represented by nodes and data sources are represented by different types of hyperedges. Each node of the unified hypergraph is embedded into a latent feature space. The unified hypergraph is partitioned into clusters within the latent feature space, where each cluster contains correlated CO data. The correlated CO data from a cluster are used to generate and provide targeted messages specific to the consumer product to a display device.

Hypergraph-based method for segmenting and clustering customer observables for vehicles

A system and method of performing fault diagnosis and analysis for one or more vehicles. The method includes: obtaining design failure mode and effect analysis (DFMEA) data that specifies a plurality of failure modes; receiving diagnostic association data; receiving vehicle operation signals association data; generating augmented DFMEA data that indicates a causal relationship between the diagnostic data and the first set of failure modes, and that indicates a causal relationship between the vehicle operation signals data and the second set of failure modes, wherein the augmented DFMEA data is generated based on the DFMEA data, the diagnostic association data, and the vehicle operation signals association data; and performing fault diagnosis and analysis for the one or more vehicles using the augmented DFMEA data.

Vehicle fault diagnosis and analysis based on augmented design failure mode and effect analysis (DFMEA) data

A system, for filtering and fusing multi-source ontologies. The system includes a tangible processing controller unit and non-transitory computer-readable storage device in communication with the tangible processing controller unit. The storage device includes a first receiving unit that, when executed by the tangible processing control unit, receives a plurality of ontologies, each ontology having a set of rules and a class structure with a plurality of data classes. The storage device also includes a second receiving unit that, when executed, receives data. The device also includes a comparison unit that compares the data classes from the plurality of ontologies, and a merging unit that merges the data classes that are identical or consistent into a new data class. The storage device also includes a discarding unit that discards the data classes that are inconsistent. The storage device also includes a new-set-generation unit that generates a new set of class structure.

Methodology for generating a consistent semantic model by filtering and fusing multi-source ontologies

A method and system to determine relatedness select a first customer observable from a first source document, the first customer observable being made up of two terms, the two terms being a first term of a first type and a first term of a second type, and select a second customer observable from a second source document, the second customer observable being made up of a second term of the first type and a second term of the second type. The method includes creating a first corpus of all documents that include the first terms, creating a second corpus of all documents that include the second terms, obtaining other first terms in the first corpus and other second in the second corpus, and performing semantic similarity analysis to determine a similarity score between the first customer observable and the second customer observable.

Semantic similarity analysis to determine relatedness of heterogeneous data

A system and method of modifying a vehicle service database includes: accessing a database containing previously-received symptom text that has been associated with a vehicle identifier and one or more vehicular service solutions for the previously-received symptom text; determining a statistical likelihood that one or more additional vehicular service solutions apply to previously-received symptom text based on a correlation between the previously-received symptom text and additional vehicular service solutions; determining that the statistical likelihood is above a predetermined threshold; and associating the previously-received symptom text with the additional vehicular service solutions.

Modifying vehicle fault diagnosis based on statistical analysis of past service inquiries

A system and method of analyzing content of multilingual vehicle diagnostic records includes: determining a word window within a vehicle diagnostic record; identifying a pair or a tuple comprising parts, symptoms, or actions; generating a plurality of pairs or tuples comprising parts, symptoms, or actions; determining a frequency value for each pair or tuple; and comparing the determined frequency value with a predetermined threshold.

Analyzing multilingual diagnostic records

A method of automatic identifying linking relationships of requirements in a plurality of requirement documents. Terms in the plurality of requirement documents are identified. A part-of-speech tag is assigned to each term. Each identified term is selected as a focal term. Co-occurring terms within a predetermined distance of the selected focal term are determined. A linking relationship probability is calculated for each co-occurring term associated with the selected focal term. The selected focal terms and associated co-occurring terms between the plurality of requirement documents are compared. A degree of linking relationship is identified between two requirements as a function of a comparison between selected focal terms and the associated co-occurring terms between the plurality of requirement documents. An analysis report identifying the degree of linking relationships between two respective requirements is output.

Automatic linking of requirements using natural language processing

A method of automatically developing an ontology for product function and failure mode documentation for an apparatus. The apparatus is identified. A function-flow model is generated for the identified apparatus for identifying a composite structure of the apparatus. Functions and failure modes associated with the identified apparatus are enumerated. Failure data is obtained from a plurality of heterogeneous data sources. A semantic similarity module is applied to the enumerated failures by comparing a plurality of documents between the data sources. The semantic similarity module bridges a variety of terms used in the heterogeneous data to describe a respective failure. Failures associated with the enumerated apparatus functions are extracted from the plurality of documents between heterogeneous data sources. A composite of related terminology is generated for each identified failure mode. A failure mode information document is generated utilizing the composite of related terminology for each identified failure mode.

Function-based method for classifying and fusing system behavior information in product development

A document may be received at a processing module. One or more tags may be applied to the document, each tag applied to a term, each tag representing a part of speech. One or more terms may be extracted from the document based on the tag. A weighting assignment parameter may be determined for each of the one or more extracted terms. Based on the weighting assignment parameter associated with each of the extracted terms, it may be determined whether the domain ontology includes the one or more extracted terms. If the domain ontology does not include the one or more extracted terms, the domain ontology may be augmented such that the domain ontology comprises the one or more extracted terms.

Method and system to augment vehicle domain ontologies for vehicle diagnosis

A method for extracting data from service repair verbatims in a vehicle service reporting system. Each service repair verbatim includes a technician's comments concerning a part, a symptom associated with the part, and a repair action associated with the symptom. Each service repair verbatim includes information relating to an identified problem with at least one vehicle part. A diagnostic and prognostic ontology database is provided that is structured by vehicle part classification, a vehicle part sub-class classification, and a relationship classification, wherein the relationship classification includes symptom relationships and action relationships. Each of the service repair verbatims are reconstructed utilizing the diagnostic and prognostic ontology database. Combinations of information are extracted from the reconstructed service repair verbatims as a function of user input criteria. A frequency is determined of each combination extracted in the reconstructed service repair verbatims. The service repair verbatims are clustered for each combination.

Knowledge extraction methodology for unstructured data using ontology-based text mining

A method and system for splitting a text document into individual sentences using sentence boundary detection, and establishing co-relationships between terms which are present in the same sentence. A document corpus, or collection of text records, is provided, containing text with terms to be extracted. The text records in the document corpus are divided into individual sentences, using a set of rules for sentence boundary detection. The individual sentences are then analyzed to extract and correlate terms, such as parts and symptoms, symptoms and actions, or parts and failure modes. The correlated terms are then validated based on frequency of occurrence, with term pairs being considered valid if their frequency of occurrence exceeds a minimum frequency threshold. The validated term correlations can be used for fault model development, document classification, and document clustering.

Growing community of inventors

Troy, MI, United States of America

Dnyanesh Rajpathak