Domain-Specific Computational Lexicon Formation via Pattern Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Natural-language processing systems face challenges in forming effective domain-specific computational lexicons from unstructured glossaries, as existing methods require extensive manual annotation and struggle with syntactic and semantic information extraction from domain-specific texts, especially in rapidly evolving domains like medicine.

Innovation Solution

A method that extracts candidate token sequences from an unstructured domain glossary, performs look-up operations to annotate them with language data, and refines these sequences through pattern matching to create a domain-specific computational lexicon, incorporating lexical information and handling ambiguous terms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If manual annotation is used to create a domain-specific computational lexicon, then the quality and completeness of syntactic and semantic information is improved, but the time and labor required increases significantly

Engineering Contradiction:

Improvequality of syntactic and semantic informationVSAvoidtime and labor for annotation

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system enables automatic self-annotation of domain-specific terms by extracting candidate token sequences from unstructured glossaries and automatically annotating them with syntactic and semantic information from external lexical resources, eliminating the need for manual annotation while maintaining quality

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of annotation with an automated computational system that uses pattern matching algorithms and external lexical resources to automatically generate syntactic and semantic annotations for domain-specific terms

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If external lexical resources are used to annotate domain-specific terms, then the speed of lexicon formation is improved, but the accuracy of syntactic and semantic information may deteriorate

Engineering Contradiction:

Improvespeed of lexicon formationVSAvoidaccuracy of syntactic and semantic information

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent introduces pattern matching as an intermediary layer between external lexical resources and domain-specific terms, which filters and adapts the retrieved information to ensure accuracy while maintaining automated processing speed

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system applies different processing strategies to different parts of the annotation process: using automated look-up for common terms while applying pattern matching refinement for domain-specific terminology, ensuring locally optimized accuracy for each type of term

Inventive Principle:

Principle #3Local quality

3Adaptability or versatility

If a comprehensive domain-specific lexicon is created to cover all domain terms, then the coverage and applicability of the NLP system is improved, but the complexity of the lexicon and processing increases

Engineering Contradiction:

Improvecoverage of domain termsVSAvoidcomplexity of lexicon structure

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the lexicon formation process into distinct modules: extraction of candidate token sequences, look-up of external lexical resources, pattern matching for syntactic analysis, and semantic annotation, allowing each module to handle specific aspects independently and reducing overall system complexity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system creates a universal pattern matching framework that can handle multiple domains and terminology types through a single unified approach, reducing the need for domain-specific customizations and simplifying the overall lexicon structure

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS9684647B2Domain-specific computational lexicon formation

Publication Date: 2017.06.20 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

According to an aspect, a candidate token sequence including one or more word tokens is extracted from an unstructured domain glossary that includes entries associated with a domain. A look-up operation is performed to retrieve language data for each word token in the candidate token sequence and annotates each word token in the candidate token sequence found by the look-up operation with corresponding retrieved language data to form an annotated sequence. A pattern match of the annotated sequence is performed relative to a repository of patterns and identifies a best matching pattern from the repository of patterns to the annotated sequence based on matching criteria. The annotated sequence is refined with lexical information associated with the best matching pattern as a refined annotated sequence. The candidate token sequence and the refined annotated sequence are output to a domain-specific computational lexicon file.