Classifying Domain Terms via ML Features

Solution Overview

Problem

Deep question answering systems face challenges in automatically categorizing large sets of domain-specific terms into entity types and entities, which is a time-consuming task for subject matter experts and requires manual effort, hindering efficient domain adaptation and taxonomy refinement.

Innovation Solution

A method that uses training sets from domain-specific question-and-answer pairs to identify distinguishing features from a corpus, enabling an automatic machine-learning classifier to classify terms into lexical answer types and entities, thereby facilitating a domain-specific lexicon for improved question answering systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual classification of domain-specific terms is performed, then classification accuracy can be maintained, but time consumption and manual effort increase significantly

Engineering Contradiction:

Improveclassification accuracyVSAvoidtime consumption

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system enables automatic self-classification of domain-specific terms by training a machine learning classifier on distinguishing features extracted from corpora. The classifier autonomously processes large sets of terms without requiring manual intervention, thereby maintaining high classification accuracy while significantly reducing time consumption and manual effort.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical classification processes with an automated machine learning system. The classifier uses distinguishing features (syntactic, lexical, contextual) to automatically categorize terms into lexical answer types and entities, substituting human expertise with an automated intelligent system that scales efficiently.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If manual classification of domain-specific terms is performed, then classification accuracy can be maintained, but manual effort and complexity increase

Engineering Contradiction:

Improveclassification accuracyVSAvoidmanual effort

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system enables automatic self-classification of domain-specific terms by training a machine learning classifier on distinguishing features extracted from corpora. The classifier autonomously processes large sets of terms without requiring manual intervention, thereby maintaining high classification accuracy while significantly reducing time consumption and manual effort.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual mechanical classification processes with an automated machine learning system. The classifier uses distinguishing features (syntactic, lexical, contextual) to automatically categorize terms into lexical answer types and entities, substituting human expertise with an automated intelligent system that scales efficiently.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If automatic classification methods are used, then processing speed improves, but classification accuracy may deteriorate

Engineering Contradiction:

Improveprocessing speedVSAvoidclassification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system performs preliminary extraction of distinguishing features from corpora before classification. By pre-identifying and storing these features (syntactic patterns, lexical characteristics, contextual relationships), the classifier can rapidly process new terms without re-analyzing the entire corpus, thus achieving both high speed and accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces manual mechanical classification processes with an automated machine learning system. The classifier uses distinguishing features (syntactic, lexical, contextual) to automatically categorize terms into lexical answer types and entities, substituting human expertise with an automated intelligent system that scales efficiently.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20180260383A1Domain-specific method for distinguishing type-denoting domain terms from entity-denoting domain terms

Publication Date: 2018.09.13 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Large lists of domain-specific terms are classified as a particular kind of linguistic object, e.g., lexical answer type T versus canonical answer E, based on features from a domain-specific corpus which have been found to distinguish between the linguistic objects. The distinguishing features can be identified in the corpus based on sets of the linguistic objects derived from question-and-answer pairs. A classifier can be trained using the distinguishing features, and the classification carried out using that classifier. The distinguishing features can include one or more syntactic features or one or more lexical features. The linguistic objects (the T and E training sets) can be extracted from the question-and-answer pairs automatically via text analysis if manually curated lists are not available. The classified terms can be included in a domain-specific lexicon which facilitates a deep question answering system to yield an answer to a question.