Text Segmentation ML for Multilingual Sentiment Labeling

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional text segmentation systems struggle with accuracy, flexibility, and efficiency when dealing with unstructured text that contains errors, varies in grammar and punctuation, and spans multiple languages, often failing to identify sentiments accurately.

Innovation Solution

A segmentation classification system using machine learning models to identify salient portions and sentiments in unstructured text, trained on unclean text with errors across various languages, employing a segmentation machine learning model and a sentiment identifying model to generate segment and sentiment labels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional rule-based systems are used to segment text according to grammar and syntax rules, then segmentation can be performed on clean and structured text, but the system breaks down when text contains spelling errors, grammar mistakes, or does not follow standard conventions

Engineering Contradiction:

Improvesegmentation accuracyVSAvoidhandling of unstructured text

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent changes the fundamental parameter of text processing from rule-based grammatical analysis to machine learning-based pattern recognition. The system is trained on clean, structured text to learn segmentation patterns, then applies these learned patterns to unstructured text with errors, allowing it to generalize beyond strict grammar rules while maintaining segmentation accuracy

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical rule-based system with a machine learning model that automatically learns segmentation rules from training data. This substitution allows the system to handle unstructured text with spelling and grammar errors by recognizing patterns rather than requiring explicit grammatical rules

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If conventional systems assign a single sentiment to a piece of text, then the system is simple to operate, but it misses sentiments expressed in the text and fails to identify multiple sentiments

Engineering Contradiction:

Improvesentiment assignment simplicityVSAvoidmissed sentiments

Core Design Contradiction:

Ease of operationVSLoss of information

Solution Approach 1:

The patent segments text into multiple portions and assigns sentiments to each segment independently, then aggregates these segment-level sentiments to produce a comprehensive text-level sentiment analysis. This allows the system to capture multiple sentiments within a single text while maintaining the simplicity of single-sentiment assignment at the segment level

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent adds a dimensional layer to sentiment analysis by operating at both the segment level and text level. Instead of assigning a single sentiment directly to the entire text, the system first analyzes sentiments at the segment dimension, then aggregates these to produce the overall text sentiment, enabling detection of multiple coexisting sentiments

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

3Adaptability or versatility

If conventional systems use multiple machine learning models to segment text in a variety of languages, then the system can handle multiple languages, but it requires large amounts of processing data for training and large storage capabilities

Engineering Contradiction:

Improvemulti-language supportVSAvoidnumber of models required

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal machine learning model that can segment text across multiple languages using a single model architecture. The model is trained on multilingual data and learns language-specific segmentation patterns internally, eliminating the need for separate models for each language and reducing both storage requirements and computational complexity

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12579371B2Using machine learning to generate segments from unstructured text and identify sentiments for each segment

Publication Date: 2026.03.17 QUALTRICS LLC

AI summary

The present disclosure relates to systems, non-transitory computer-readable media, and methods for segmenting unstructured text into salient portions and identifying sentiments expressed in each segment. In particular, the disclosed systems utilize a segmentation machine learning model to segment unstructured text into salient portions and a sentiment identifying machine learning model to identify sentiments for each segment. Additionally, the disclosed systems determine a topic for each segment and associate it with an emotion label, a sentiment label or a predicted action label. In one or more embodiments, based on the topic associated with the emotional label, sentiment label or predicted action label, the disclosed systems determine and perform additional actions.