Back to results
US 12,651,017 B2Granted
Techniques for information ranking and retrieval
AssigneeState Street Corporation
InventorsDushyant Ralhan, Gubo Huang, Yuan Fang
Life of this patent
Estimating the term…
Abstract
Systems, methods, apparatuses, and computer-readable media for information ranking and retrieval are described. In one embodiment, an apparatus may include a processor and a memory storing instructions which when executed by the processor cause the processor to access an ingested document, generate a converted document from the ingested document based on a conversion configuration, the converted document comprising at least one paragraph, and generate an index based on the converted document and an index configuration. Other embodiments are described.
Claims11
An apparatus, comprising:
- a processor for executing an enterprise-level information retrieval system that is enterprise wide;
- and
- a memory storing instructions of the enterprise-level information retrieval system which when executed by the processor cause the processor to:
- batch upload documents, including documents containing paragraphs of text;
- ingest the documents into the enterprise-level information retrieval system, wherein as part of ingesting the documents at least one document is converted into a converted document;
- access the ingested documents;
- generate a search index for the ingested documents, according to an index configuration;
- store the ingested documents and the search index in a storage;
- generate an operating environment for information rank and retrieval processes to retrieve and rank via the rank model, wherein the operating environment comprises a search platform cloud for providing a search capability that encompasses information during document ingestion and index generation and an administrative user interface (UI) to facilitate onboarding of applications;
- generate, in the operating environment via a UI template image and based on environment configurations, containers for application specific client UIs for the information rank and retrieval processes;
- and
- retrieve and rank, via the rank model, at least one of the ingested documents responsive to a query, the rank model comprising a machine learning model.
9 more elements
A method, comprising:
- with an enterprise-level information retrieval system that is enterprise wide, batch uploading documents, including documents containing paragraphs of text;
- with the enterprise-level information retrieval system, ingesting the documents into the enterprise-level information retrieval system, wherein as part of the ingesting the documents at least one document is converted into a converted document;
- with the enterprise-level information retrieval system, accessing the ingested documents;
- with the enterprise-level information retrieval system, generating a search index for the ingested documents based on an index configuration;
- with the enterprise-level information retrieval system, store the ingested documents and search index in a storage;
- with the enterprise-level information retrieval system, generating an operating environment for information rank and retrieval processes to retrieve and rank via the rank model, wherein the operating environment comprises a search platform cloud for providing a search capability that encompasses information during document ingestion and index generation and an administrative user interface (UI) to facilitate onboarding of applications;
- with the enterprise-level information retrieval system, generating, in the operating environment via a UI template image and based on environment configurations, containers for application specific client UIs for the information rank and retrieval processes;
- and
- retrieving and ranking, via the rank model, at least one of the ingested documents responsive to a query, the rank model comprising a machine learning model.
6 more elements
A computer-readable storage medium, storing instructions for an enterprise-level information retrieval system that when executed by the processor circuit cause the processor circuit to:
- with an enterprise-level information retrieval system that is enterprise wide, batch upload documents, including documents containing paragraphs of text;
- with the enterprise-level information retrieval system, ingest the documents into the enterprise information retrieval system, wherein as part of the ingesting the documents at least one document is converted into a converted document;
- with the enterprise-level information retrieval system, access the ingested document;
- with the enterprise-level information retrieval system, generate a search index for the ingested documents based on an index configuration;
- with the enterprise-level information retrieval system, store the ingested documents and search index in a storage;
- with the enterprise-level information retrieval system, generate an operating environment for information rank and retrieval processes to retrieve and rank via the rank model, wherein the operating environment comprises a search platform cloud for providing a search capability that encompasses information during document ingestion and index generation and an administrative user interface (UI) to facilitate onboarding of applications;
- with the enterprise-level information retrieval system, generate, in the operating environment via a UI template image and based on environment configurations, containers for application specific client UIs for the information rank and retrieval processes;
- and
- retrieve and rank, via the rank model, at least one of the ingested documents responsive to a query, the rank model comprising a machine learning model.
6 more elements
Dependencies are inferred from the claim text.