Mathematical Function Annotation for Numerical Text in NLP

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing natural language annotation techniques lack adequate tools for accurately annotating numerical text within documentation, leading to poor results when numerical statements appear together or are dispersed, as they fail to provide sufficient context and computational value for machine learning models.

Innovation Solution

A mathematical natural language annotation system that identifies numerical text in samples, allows annotators to select and combine mathematical functions from a library to represent context and computational value, and inserts these functions into feature vectors for improved machine learning model training.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing natural language annotation techniques are used, then the annotation process is simple, but the accuracy and computational value for numerical text is insufficient

Engineering Contradiction:

Improveannotation accuracyVSAvoidannotation system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces mathematical functions as an intermediary layer between numerical text and machine learning models. These functions serve as mediators that transform raw numerical data into computationally meaningful representations, thereby improving annotation accuracy without requiring complex manual intervention. The mathematical library acts as an intermediary resource that automatically provides contextual information.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system enables self-service annotation by automatically identifying numerical text and suggesting appropriate mathematical functions from the library. The annotator tool autonomously performs tasks such as detecting numerical patterns, retrieving relevant mathematical functions, and inserting them into feature vectors, reducing the need for manual annotation effort while maintaining high accuracy.

Inventive Principle:

Principle #25Self-service

2Loss of information

If mathematical functions are added to provide context, then computational value improves, but annotation process complexity increases

Engineering Contradiction:

Improvecontext information retentionVSAvoidannotation tool complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The mathematical library is pre-populated with relevant mathematical functions and their metadata before the annotation process begins. This preliminary preparation allows the system to quickly retrieve and apply appropriate functions during annotation without performing complex real-time analysis, thereby preserving context information while minimizing added complexity.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The mathematical library serves multiple functions: it stores mathematical functions, provides contextual information, suggests appropriate functions for numerical text, and integrates with the annotator tool. This multi-functionality consolidates what could be separate complex systems into a single unified component, reducing overall system complexity while maintaining rich contextual information.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Reliability

If numerical text is annotated with mathematical functions, then machine learning model performance improves, but processing time increases

Engineering Contradiction:

Improvemodel training reliabilityVSAvoidannotation processing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system uses lightweight, simple mathematical function representations that are computationally inexpensive to process. These function annotations are designed to be processed quickly by machine learning models without requiring extensive computation, thereby maintaining model training reliability while minimizing additional processing time.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

Data Source

PatentUS12197846B2Mathematical function defined natural language annotation

Publication Date: 2025.01.14 INTERNATIONAL BUSINESS MACHINE CORPORATION

AI summary

Provided is a method, a computer program product, and a system for associating mathematical functions to numerical text in a natural language sample. The method includes inputting a natural language sample from a text dataset and identifying a numerical text within the natural language sample. The method further includes displaying a mathematical function corresponding to the numerical text to be selected. The mathematical function can be selected via graphical user interface displayed on a computing device. The method also includes receiving and inserting the mathematical function as a feature into a feature vector of the natural language sample and selecting an output label for the natural language sample. The output label relates to the mathematical function selected for the numerical text. The method further includes exporting the natural language sample into a labeled dataset which can be used to train a machine learning model.