Patents by Inventor Sanjay BODDHU
Sanjay BODDHU has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Publication number: 20230136672Abstract: A model management system performs error analysis on results predicted by a machine learning model. The model management system identifies an incorrectly classified image outputted from a machine learning model and identifies using the Neural Template Matching (NTM) algorithm, an additional image correlated to the selected image. The system outputs correlated images based on a given image and a selection by a user through a user interface of a region of interest (ROI) of the given image. The region is defined by a bounding polygon input and the correlated images include features correlated to the features within the ROI. The system prompts a task associated with the additional image. The system receives a response that includes an indication that the additional image is incorrectly labeled and including a replacement label and instruct that the machine learning model be retrained using an updated training dataset that includes the replacement label.Type: ApplicationFiled: October 21, 2022Publication date: May 4, 2023Inventors: Mark William Sabini, Kai Yang, Andrew Yan-Tak Ng, Daniel Bibireata, Dillon Laird, Whitney Blodgett, Yan Liu, Yazhou Cao, Yuxiang Zhang, Gregory Diamos, YuQing Zhou, Sanjay Boddhu, Quinn Killough, Shankaranand Jagadeesan, Camilo Zapata, Sebastian Rodriguez
-
Patent number: 11189032Abstract: A method, apparatus and computer program product are provided for extracting building footprints for buildings based on satellite imagery. In this regard, a segmentation map is generated by performing an image segmentation on satellite imagery depicting at least one building. A plurality of line segments that correspond to edges of the at least one building are also generated. Furthermore, a contour for each of the at least one building is generated by filtering line segments corresponding to the at least one building based on the respective principle direction. The geographic data for each of the at least one building is then extracted based on the respective contour.Type: GrantFiled: April 1, 2020Date of Patent: November 30, 2021Assignee: HERE GLOBAL B.V.Inventors: Di Ma, Lin Gan, Fangzhou Xiong, Sanjay Boddhu
-
Publication number: 20210312632Abstract: A method, apparatus and computer program product are provided for extracting building footprints for buildings based on satellite imagery. In this regard, a segmentation map is generated by performing an image segmentation on satellite imagery depicting at least one building. A plurality of line segments that correspond to edges of the at least one building are also generated. Furthermore, a contour for each of the at least one building is generated by filtering line segments corresponding to the at least one building based on the respective principle direction. The geographic data for each of the at least one building is then extracted based on the respective contour.Type: ApplicationFiled: April 1, 2020Publication date: October 7, 2021Inventors: Di Ma, Lin Gan, Fangzhou Xiong, Sanjay Boddhu
-
Patent number: 9984427Abstract: A system and method for detecting and summarizing events based on data feeds from a plurality of sources. Such sources may include social media networks, text messages, news feeds among others. The system may receive raw information from such sources containing data related with possible events. Method for event detection may include pre-processing and normalizing data input from any source registered, this may also include; extracting and disambiguating events and entities, associate event and entities, correlate events and entities associated from a data input which results from a different data source, for validating/verifying an event. Subsequently, the validated/verified event may be stored in a local data storage and/or in a web-server.Type: GrantFiled: December 2, 2014Date of Patent: May 29, 2018Assignee: QBase, LLCInventors: Rakesh Dave, Sanjay Boddhu
-
Patent number: 9922032Abstract: A system for building a knowledge base of co-occurring features extracted from a document corpus is disclosed. The method includes a plurality of feature extraction software modules that may extract different features from each document in the corpus. The system may include a knowledge base aggregator module that may keep count of the co-occurrences of features in the different documents of a corpus and determine appropriate co-occurrences to store in a knowledge base.Type: GrantFiled: December 2, 2014Date of Patent: March 20, 2018Assignee: QBase, LLCInventors: Scott Lightner, Rakesh Dave, Sanjay Boddhu
-
Patent number: 9910723Abstract: A system and method for detecting events based on input data from a plurality of sources. The system may receive input from a plurality of sources containing information about possible events. A method for event detection involves pre-processing and normalizing a data input from a plurality of sources, extracting and disambiguating events and entities, associate event and entities, correlate events and entities associated from a data input to results from a different data source to determine if an event has occurred, and store the detected events in a data storage.Type: GrantFiled: January 9, 2017Date of Patent: March 6, 2018Assignee: QBase, LLCInventors: Sanjay Boddhu, Rakesh Dave
-
Publication number: 20170286837Abstract: The present disclosure relates to a method for performing automated discovery of new topics from unlimited documents related to any subject domain, employing a multi-component extension of Latent Dirichlet Allocation (MC-LDA) topic models, to discover related topics in a corpus. The resulting data may contain millions of term vectors from any subject domain identifying the most distinguished co-occurring topics that users may be interested in, for periodically building new topic ID models using new content, which may be employed to compare one by one with existing model to measure the significance of changes, using term vectors differences with no correlation with a Periodic New Model, for periodic updates of automated discovery of new topics, which may be used to build a new topic ID model in-memory database to allow query-time linking on massive data-set for automated discovery of new topics.Type: ApplicationFiled: April 17, 2017Publication date: October 5, 2017Inventors: Scott LIGHTNER, Franz WECKESSER, Sanjay BODDHU, Robert FLAGG
-
Publication number: 20170212899Abstract: A method for searching for related entities using entity co-occurrence is disclosed. Embodiments of the method may be employed in any search system that may include at least one search engine, at least one entity co-occurrence knowledge base, an entity extraction module, and at least an entity indexed corpus. The method may extract and disambiguate entities from search queries by using an entity co-occurrence knowledge base, find extracted entities in an entity indexed corpus and finally present search results as related entities of interest.Type: ApplicationFiled: April 7, 2017Publication date: July 27, 2017Inventors: Scott LIGHTNER, Franz Weckesser, Sanjay Boddhu
-
Publication number: 20170199914Abstract: A method for generating search suggestions of related entities based on co-occurrence and/or fuzzy score matching is disclosed. The method may be employed in a search system that may include a client/server type architecture. The search system may include a user interface for a search engine in communication with one or more server devices over a network connection. The server device may include an entity extraction module, a fuzzy-score matching module, and an entity co-occurrence knowledge base database. In one embodiment, the search system may process a partial search query from a user and present search suggestions to complete the partial query. In another embodiment, the complete search query may be used as a new search query. The search system may process the new search query, run an entity extraction, find related entities from the entity co-occurrence knowledge base, and present said related entities in a drop down list.Type: ApplicationFiled: March 27, 2017Publication date: July 13, 2017Inventors: Scott LIGHTNER, Franz WECKESSER, Sanjay BODDHU
-
Publication number: 20170124090Abstract: Methods and systems for discovering and exploring feature knowledge included in large corpora are disclosed. The described systems and methods may include the application of in-memory analytics to records, where the analytic methods applied to the records and the level of precision of the methods may be dynamically selected by a user.Type: ApplicationFiled: January 13, 2017Publication date: May 4, 2017Inventors: Scott LIGHTNER, Sanjay BODDHU, Robert FLAGG
-
Publication number: 20170116203Abstract: A computer system and method for automated discovery of topic relatedness are disclosed. According to an embodiment, topics within documents from a corpus may be discovered by applying multiple topic identification (ID) models, such as multi-component latent Dirichlet allocation (MC-LDA) or similar methods. Each topic model may differ in a number of topics. Discovered topics may be linked to the associated document. Relatedness between discovered topics may be determined by analyzing co-occurring topic IDs from the different models, assigning topic relatedness scores, where related topics may be used for matching/linking a feature of interest. The disclosed method may have an increased disambiguation precision, and may allow the matching and linking of documents using the discovered relationships.Type: ApplicationFiled: January 9, 2017Publication date: April 27, 2017Inventors: Scott LIGHTNER, Franz WECKESSER, Sanjay BODDHU, Rakesh DAVE, Robert FLAGG
-
Publication number: 20170116054Abstract: A system and method for detecting events based on input data from a plurality of sources. The system may receive input from a plurality of sources containing information about possible events. A method for event detection involves pre-processing and normalizing a data input from a plurality of sources, extracting and disambiguating events and entities, associate event and entities, correlate events and entities associated from a data input to results from a different data source to determine if an event has occurred, and store the detected events in a data storage.Type: ApplicationFiled: January 9, 2017Publication date: April 27, 2017Inventors: Sanjay BODDHU, Robert FLAGG, Rakesh DAVE, Scott LIGHTNER
-
Patent number: 9626623Abstract: The present disclosure relates to a method for performing automated discovery of new topics from unlimited documents related to any subject domain, employing a multi-component extension of Latent Dirichlet Allocation (MC-LDA) topic models, to discover related topics in a corpus. The resulting data may contain millions of term vectors from any subject domain identifying the most distinguished co-occurring topics that users may be interested in, for periodically building new topic ID models using new content, which may be employed to compare one by one with existing model to measure the significance of changes, using term vectors differences with no correlation with a Periodic New Model, for periodic updates of automated discovery of new topics, which may be used to build a new topic ID model in-memory database to allow query-time linking on massive data-set for automated discovery of new topics.Type: GrantFiled: October 21, 2015Date of Patent: April 18, 2017Assignee: QBase, LLCInventors: Scott Lightner, Franz Weckesser, Sanjay Boddhu, Robert Flagg
-
Patent number: 9619571Abstract: A method for searching for related entities using entity co-occurrence is disclosed. Embodiments of the method may be employed in any search system that may include at least one search engine, at least one entity co-occurrence knowledge base, an entity extraction module, and at least an entity indexed corpus. The method may extract and disambiguate entities from search queries by using an entity co-occurrence knowledge base, find extracted entities in an entity indexed corpus and finally present search results as related entities of interest.Type: GrantFiled: December 2, 2014Date of Patent: April 11, 2017Assignee: QBase, LLCInventors: Scott Lightner, Franz Weckesser, Sanjay Boddhu
-
Patent number: 9613166Abstract: A method for generating search suggestions of related entities based on co-occurrence and/or fuzzy score matching is disclosed. The method may be employed in a search system that may include a client/server type architecture. The search system may include a user interface for a search engine in communication with one or more server devices over a network connection. The server device may include an entity extraction module, a fuzzy-score matching module, and an entity co-occurrence knowledge base database. In one embodiment, the search system may process a partial search query from a user and present search suggestions to complete the partial query. In another embodiment, the complete search query may be used as a new search query. The search system may process the new search query, run an entity extraction, find related entities from the entity co-occurrence knowledge base, and present said related entities in a drop down list.Type: GrantFiled: October 22, 2015Date of Patent: April 4, 2017Assignee: QBase, LLCInventors: Scott Lightner, Franz Weckesser, Sanjay Boddhu
-
Publication number: 20170075915Abstract: A method for generating search suggestions by using fuzzy-score matching and entity co-occurrence in a knowledge base is disclosed. Embodiments of the method may be employed in any search system that may include an entity extraction computer module that may perform partial entity extractions from provided search queries, a fuzzy-score matching computer module that may generate algorithms based on the type of entity extracted and perform a search against an entity co-occurrence knowledge base. The entity co-occurrence knowledge base, which may include a repository where entities may be indexed as entities to entities, entities to topics, or entities to facts among others, may return fast and accurate suggestions to the user to complete the search query. The suggestions may include alternates to the partial query provided by the user that may enhance and save time when performing searches.Type: ApplicationFiled: November 28, 2016Publication date: March 16, 2017Inventors: Scott LIGHTNER, Franz WECKESSER, Rakesh DAVE, Sanjay BODDHU, Joseph BECKNELL
-
Patent number: 9547701Abstract: Methods and systems for discovering and exploring feature knowledge included in large corpora are disclosed. The described systems and methods may include the application of in-memory analytics to records, where the analytic methods applied to the records and the level of precision of the methods may be dynamically selected by a user.Type: GrantFiled: December 2, 2014Date of Patent: January 17, 2017Assignee: QBase, LLCInventors: Scott Lightner, Sanjay Boddhu, Robert Flagg
-
Patent number: 9544361Abstract: A system and method for detecting events based on input data from a plurality of sources. The system may receive input from a plurality of sources containing information about possible events. A method for event detection involves pre-processing and normalizing a data input from a plurality of sources, extracting and disambiguating events and entities, associate event and entities, correlate events and entities associated from a data input to results from a different data source to determine if an event has occurred, and store the detected events in a data storage.Type: GrantFiled: December 2, 2014Date of Patent: January 10, 2017Assignee: QBase, LLCInventors: Sanjay Boddhu, Rakesh Dave
-
Patent number: 9542477Abstract: A computer system and method for automated discovery of topic relatedness are disclosed. According to an embodiment, topics within documents from a corpus may be discovered by applying multiple topic identification (ID) models, such as multi-component latent Dirichlet allocation (MC-LDA) or similar methods. Each topic model may differ in a number of topics. Discovered topics may be linked to the associated document. Relatedness between discovered topics may be determined by analyzing co-occurring topic IDs from the different models, assigning topic relatedness scores, where related topics may be used for matching/linking a feature of interest. The disclosed method may have an increased disambiguation precision, and may allow the matching and linking of documents using the discovered relationships.Type: GrantFiled: December 2, 2014Date of Patent: January 10, 2017Assignee: QBase, LLCInventors: Scott Lightner, Franz Weckesser, Sanjay Boddhu, Rakesh Dave, Robert Flagg
-
Patent number: 9507834Abstract: A method for generating search suggestions by using fuzzy-score matching and entity co-occurrence in a knowledge base is disclosed. Embodiments of the method may be employed in any search system that may include an entity extraction computer module that may perform partial entity extractions from provided search queries, a fuzzy-score matching computer module that may generate algorithms based on the type of entity extracted and perform a search against an entity co-occurrence knowledge base. The entity co-occurrence knowledge base, which may include a repository where entities may be indexed as entities to entities, entities to topics, or entities to facts among others, may return fast and accurate suggestions to the user to complete the search query. The suggestions may include alternates to the partial query provided by the user that may enhance and save time when performing searches.Type: GrantFiled: November 24, 2015Date of Patent: November 29, 2016Assignee: QBase, LLCInventors: Scott Lightner, Franz Weckesser, Rakesh Dave, Sanjay Boddhu, Joseph Becknell