Synonym-Sensitive Unit Conversion for QA Measurement Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current QA systems are unable to provide complete and accurate answers to questions that include measurement units or values due to their inability to handle different measurement units, such as converting between 'feet' and 'meters' or 'inches', leading to incomplete results.

Innovation Solution

The implementation of a synonym-sensitive unit conversion framework within the QA system, which analyzes questions for measurement values, converts these values to a default or specified unit, and assigns confidence scores based on proximity, enabling accurate comparison and ranking of relevant passages.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the QA system uses exact string matching for measurement values, then the comparison is simple and fast, but the system cannot recognize equivalent values in different units (e.g., '7 feet' vs. '2.13 meters')

Engineering Contradiction:

Improveunit recognition capabilityVSAvoidmeasurement comparison mechanism

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a measurement value normalization mechanism that converts different unit representations into a common standardized form. This intermediary normalization step enables the system to recognize equivalent measurement values across different units without requiring complex direct comparison logic between all possible unit pairs.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the parameter representation of measurement values by converting them into a normalized standard form. This parameter transformation allows equivalent measurements (e.g., 7 feet and 2.13 meters) to be represented identically after normalization, enabling simple equality comparison while maintaining unit versatility.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If the QA system converts all measurement values to a default unit, then accurate comparisons can be made, but the system loses information about the original units specified in the question

Engineering Contradiction:

Improvemeasurement value comparison accuracyVSAvoidoriginal unit information

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The system performs preliminary normalization of measurement values to a standard form before comparison operations, preserving the ability to accurately compare equivalent measurements. The original unit information is maintained and associated with the normalized value, allowing the system to restore or reference the original units after comparison if needed.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates a normalized copy of the measurement value for comparison purposes while retaining the original value with its unit information. This copying approach allows accurate comparison through the normalized version while preserving the original unit data for reference or output.

Inventive Principle:

Principle #26Copying

3Reliability

If the QA system performs unit conversion for all relevant passages, then complete and accurate answers can be provided, but the processing time and computational resources increase

Engineering Contradiction:

Improveanswer completenessVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system applies measurement normalization selectively rather than universally. It performs exact matching first for efficiency, and only applies the more computationally intensive normalization process when exact matches fail or when unit differences are detected. This partial application of the conversion process maintains answer completeness while reducing overall processing time.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent segments the measurement comparison process into distinct stages: initial exact string matching, followed by normalized comparison only for candidates that fail the first stage. This segmentation allows the system to benefit from the speed of simple matching while still achieving complete and accurate results through the more thorough normalization process when needed.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11544605B2Unit conversion in a synonym-sensitive framework for question answering

Publication Date: 2023.01.03 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

A question and answer (QA) system, computer program product, and computer-implemented method configured to determine an answer to a question that includes a measurement value. In one example, the QA system receives a question and analyzes the question to identify a measurement value specified in the question. The QA system determines relevant passages to the question. The QA system assigns a measurement value confidence score to a relevant passage based on a comparison of the measurement value specified in the question and a second measurement value specified in the relevant passage. The QA system determines an order of the relevant passages using the measurement value confidence score of each of the relevant passages. The QA system determines an answer to the question based on the order of the relevant passages.