Finite State Transducers for Text Disambiguation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Natural language processing technologies face challenges in accurately interpreting and disambiguating words with multiple meanings, leading to inefficient and inaccurate automated text analysis.

Innovation Solution

The use of finite state transducers (FST) and bi-machine transducers (BMT) is applied to language space data to disambiguate word meanings and enhance sentence structure interpretation, allowing for the generation of modified language space data that improves the accuracy and speed of automated text analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional natural language processing algorithms are used to interpret text, then the system can process text at all, but the accuracy of word meaning disambiguation is poor and computational speed is slow

Engineering Contradiction:

Improveword meaning disambiguation accuracyVSAvoidcomputational processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies finite state transducers and bi-machine transducers to pre-process language space data before full NLP analysis. By performing preliminary disambiguation and structural analysis using these specialized transducers, the system resolves word meanings earlier in the processing pipeline, reducing the computational burden and time required for subsequent analysis steps.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces finite state transducers and bi-machine transducers as intermediary processing layers between raw text input and final NLP interpretation. These transducers act as mediators that transform and structure the language space data, enabling more accurate and efficient disambiguation of word meanings while maintaining processing speed.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If computationally intensive algorithms are used to decipher ambiguous text, then word meaning accuracy may improve, but processing speed decreases significantly

Engineering Contradiction:

Improvetext interpretation accuracyVSAvoidtext processing throughput

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent segments the text processing task into distinct phases handled by different computational mechanisms. Finite state transducers handle specific disambiguation patterns, bi-machine transducers handle structural analysis, and conventional NLP algorithms handle remaining interpretation. This segmentation allows each component to operate efficiently on its designated subtask, maintaining high accuracy without requiring all algorithms to run at full computational intensity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the computational parameters by using finite state and bi-machine transducers that operate with different complexity characteristics than conventional NLP algorithms. These transducers use optimized state transitions and matching mechanisms that reduce computational overhead while maintaining or improving disambiguation accuracy, thereby increasing overall processing throughput.

Inventive Principle:

Principle #35Parameter changes

3Productivity

If simple processing methods are used, then processing speed is fast, but the ability to correctly interpret words with multiple meanings is lost

Engineering Contradiction:

Improveprocessing speedVSAvoidword meaning disambiguation accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

Inventive Principle:

Principle #24Intermediary (Mediator)

4Ease of operation

If conventional language processors are used, then text can be processed, but sentences requiring complex interpretation require excessive computational resources

Engineering Contradiction:

Improveautomated text analysis capabilityVSAvoidcomputational resource requirements

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11568150B2Methods and apparatus to improve disambiguation and interpretation in automated text analysis using transducers applied on a structured language space

Publication Date: 2023.01.31 CLOVER AI LLC

AI summary

Methods and apparatus for automated processing of natural language text is described. The text can be preprocessed to produce language-space data that includes descriptive data elements for words. Source code that includes linguistic expressions, and that may be written in a programming language that is user-friendly to linguists, can be compiled to produce finite-state transducers and bi-machine transducers that may be applied directly to the language-space data by a language-processing virtual machine. The language-processing virtual machine can select and execute code segments identified in the finite-state and/or bi-machine transducers to disambiguate meanings of words in the text.