Knowledge management system and process for managing knowledge
Inventors
Elliott, John • Bhat, Talapady N. • Kattner, Ursula R. • Campbell, Carelyn E. • Sriram, Ram D. • Subrahmanian, Eswaran • Collard, Jacob • Monarch, Ira
Assignees
United States Department of Commerce
Publication Number
US-10872122-B2
Publication Date
2020-12-22
Expiration Date
2038-01-30
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
A knowledge management system includes: a default knowledge system including: a knowledge system and a knowledge database in communication with the knowledge system; and a knowledge store in communication with the default knowledge system and including: a taxonomy amendment, an annotation amendment, a canonicalization amendment, an ecosystem amendment, a term amendment, and a phrase amendment.
Core Innovation
The invention discloses a knowledge management system and processes for managing knowledge that use a root- and rule-based methodology combined with natural language processing tools to index documents and create phrase-based structured keywords encoding semantics. This system facilitates knowledge search by considering terms and their context to disambiguate results, continuously evolving terminology within a domain, and generating deterministic and consistent domain-based taxonomies and ecosystems of concepts that evolve during search and discovery processes.
The knowledge management system includes a default knowledge system with a knowledge system and knowledge database, and a knowledge store containing various amendment stores such as taxonomy, annotation, canonicalization, ecosystem, term, and phrase amendments. The system modularizes documents into objects, sentences, snippets, words, and canonicalized roots to reduce linguistic variability such as synonyms and jargon. It converts canonicalized words into phrases and terms to create use-case specific taxonomies and ecosystems representing relationships between knowledge components.
The problem addressed by the invention is the complexity of managing large collections of natural language documents in diverse domains, where conventional combinatorial methods and manual human processes are inadequate. The invention provides a unified, adaptable framework that produces structured, semantically meaningful representations of terms, taxonomies, and concept ecologies, supports query elaboration and discovery in technical and scientific document collections, and can be fine-tuned by domain-specific amendments to improve knowledge management efficiency and effectiveness.
Claims Coverage
The patent presents multiple independent claims related to processes, computer-implemented methods, systems, and computer-readable media for managing knowledge, each focusing on orchestrated communications between knowledge components and use of algebraic data types for phrase annotation.
Coordinated knowledge management process involving multiple managers
The invention features a process in which a knowledge manager requests documents, taxonomies, ecosystems, and annotations from respective managers or databases. The annotation includes an algebraic data type for phrases with a structure comprising a leaf with a lexical string, left and right subtrees, a head field indicating the syntactic head, and linguistic features representing syntactic category and dependency types. The knowledge manager communicates integrated knowledge management data to a knowledge warehouse.
Taxonomy creation with amendment integration
The process includes taxonomy creation by a taxonomy manager, builder, and framework structured to request and communicate taxonomy schemes and amendments from a knowledge store, enabling customization and refinement of taxonomy building.
Ecosystem management with scheme and amendment incorporation
Ecosystem generation involves requesting ecosystem snippets and schemes by ecosystem manager, generator, and framework with provisions for requesting and incorporating ecosystem amendments from a knowledge store.
Annotation scheme management with amendments
Annotation manager and framework coordinate requests and communication of annotation scheme and amendments from a knowledge store to enable domain-specific control over document annotations.
Term generation scheme with amendment support
Term generation is controlled by an annotation manager, term generator, and term framework which request, communicate, and incorporate term generation schemes and amendments from a knowledge store.
Phrase generation and canonicalization with amendment inclusion
Phrase generator and phrase canonicalizer operate with requests and communications of phrase generation schemes and canonicalization schemes, including receiving amendments from a knowledge store to ensure flexible and domain-specific phrase canonicalization.
Computer-implemented methods and systems mirroring the process
The invention also covers computer-implemented methods, systems with one or more computers, and non-transitory computer-readable media executing operations as described in the process claims, including the structured handling of document, taxonomy, ecosystem, and annotation data with phrase algebraic data types.
The claims cover a comprehensive and modular knowledge management framework implementing a root- and rule-based knowledge system with natural language processing, wherein documents are handled through coordinated management components that request, process, amend, and communicate structured knowledge data including documents, taxonomies, ecosystems, annotations, terms, and phrases represented by an algebraic data type embodying linguistic structure.
Stated Advantages
The system creates phrase-based structured keywords that encode semantics and facilitate refined searches with contextual disambiguation.
It supports continuous incremental evolution of terminology and domain-based taxonomies and ecosystems.
The approach deterministically and consistently builds concept taxonomies and ecologies, enabling knowledge evolution during search and discovery.
It provides computational efficiency not achievable by humans alone, managing complexity with time complexities ranging from O(n²) to O(m) depending on document and system parameters.
The integration of machine learning with formal computational linguistic models allows adaptable amendments and updates to account for new knowledge or data.
Documented Applications
Managing and indexing large technical and scientific document collections to support search, query elaboration, keyword selection, taxonomy, and concept ecology building.
Detecting, tracking, and combining ideas across scientific and technical domains and economic markets to develop novel concepts for research and technology creation.
Transforming linguistic outputs into structured representations useful for information retrieval, topic modeling, and domain-specific knowledge management.
Interested in licensing this patent?