Contraction Expansion Using Word Movers Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine-based methods for expanding contractions in text are often inaccurate due to the need for contextual knowledge and the presence of multiple possible expansion forms, leading to unreliable results in electronically generated and edited text.
Innovation Solution
A three-pass method that identifies simple contractions, applies grammar checking, and uses Word Movers Distance calculations to select the most grammatically correct and contextually appropriate expansion form by converting text into vector representations and calculating the distance between original and expanded forms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If simple replacement rules are used to expand contractions, then the process is simple and fast, but the accuracy is low due to multiple possible expansion forms
Solution Approach 1:
The patent segments the contraction expansion process into three distinct passes: (1) identifying contractions with single expanded forms using simple rules, (2) applying grammar checking to evaluate multiple expansion forms, and (3) using Word Movers Distance calculations for ambiguous cases. This segmentation allows the system to apply appropriate complexity only where needed, resolving the contradiction between accuracy and system complexity.
Solution Approach 2:
The patent changes the parameter of evaluation from simple pattern matching to grammar-based scoring and vector distance metrics. By introducing grammar scores and Word Movers Distance calculations as additional parameters, the system achieves higher accuracy in disambiguating contraction expansions without requiring complete system redesign.
2Measurement precision
If grammar checking is applied to evaluate multiple expansion forms, then the accuracy improves, but the processing time increases
Solution Approach 1:
The patent applies partial action by using grammar checking only for contractions that have multiple possible expansion forms, while using simple replacement rules for contractions with single expanded forms. This selective application of complex processing reduces overall processing time while maintaining high accuracy for ambiguous cases.
Solution Approach 2:
Different quality levels of processing are applied to different types of contractions: simple contractions receive basic replacement, while ambiguous contractions receive full grammar checking and Word Movers Distance analysis. This local differentiation optimizes the balance between processing time and accuracy.
3Measurement precision
If Word Movers Distance calculations are used to resolve ambiguous contractions, then the contextual accuracy improves, but the computational complexity increases
Solution Approach 1:
The patent performs preliminary action by converting text to vector representations using word embeddings before the actual distance calculation. This preprocessing step enables more efficient computation of Word Movers Distance by having the vector space ready in advance, reducing the computational burden during the actual contraction resolution process.
Solution Approach 2:
The patent introduces word embeddings as an intermediary representation between the original text and the distance calculation. This intermediary vector space simplifies the computational complexity of comparing contraction expansions by transforming the problem into vector space operations rather than direct text comparison.
Data Source
AI summary
As described herein, a system for expanding contractions in electronically stored text includes expanding contractions having only on expanded form. For remaining contractions, a grammar check is performed for all possible expanded forms to determine if an expanded form can be selected based on context and grammar rules. If an expanded form is not evident from the first two steps, all possible expanded forms of the remaining contractions are converted to a vector representation along with the original text. A Word Movers Distance (WMD) for each possible expansion is calculated using the vectors for each possible expansion and the original text. An expanded form is chosen without human intervention based on the grammar score alone or the WMD and the grammar score.
