Text Waveform Analysis for Multilingual Anomalous Message Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems face challenges in handling a wide variety of languages and languages in real-time, particularly in detecting malicious messages such as spam and viruses, due to the need for extensive data and complex analysis, which increases cost and reduces processing speed.
Innovation Solution
A text analysis system that converts text data into a time series signal, extracts features using an auto-encoder neural network, and determines anomalies based on these features, without requiring morphological analysis or language-specific dictionaries, allowing for low-cost and fast detection across multiple languages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If natural language understanding is used to detect malicious messages, then detection accuracy is improved, but processing speed deteriorates and system complexity increases
Solution Approach 1:
The patent extracts only the essential features needed for spam detection from complete natural language understanding. Instead of performing full morphological analysis and semantic interpretation, the system extracts key linguistic features (such as word frequency, character patterns, and structural characteristics) that are sufficient for detecting malicious messages, thereby reducing processing complexity while maintaining detection accuracy.
Solution Approach 2:
The patent segments the natural language processing task into distinct feature extraction components rather than performing complete language understanding. The system divides text analysis into separate feature detection modules that independently extract relevant characteristics, avoiding the computational burden of full syntactic and semantic analysis while preserving detection effectiveness.
2Measurement precision
If complete natural language understanding is performed, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the essential features needed for spam detection from complete natural language understanding. Instead of performing full morphological analysis and semantic interpretation, the system extracts key linguistic features (such as word frequency, character patterns, and structural characteristics) that are sufficient for detecting malicious messages, thereby reducing processing complexity while maintaining detection accuracy.
Solution Approach 2:
The patent inverts the traditional approach by not starting with complete language understanding and then filtering, but rather by directly extracting relevant features from raw text. This inversion simplifies the system architecture by eliminating the need for complex language understanding modules while focusing computational resources on extracting only the features necessary for detection.
3Measurement precision
If language-specific dictionaries and morphological analysis are used, then detection accuracy is improved, but processing speed and adaptability to multiple languages deteriorate
Solution Approach 1:
The patent implements a universal feature extraction approach that works across multiple languages without requiring language-specific dictionaries or morphological analysis rules. The system extracts linguistic features (such as character n-grams, word frequency patterns, and structural characteristics) that are language-agnostic, enabling the same detection mechanism to effectively process texts in various languages while maintaining both accuracy and adaptability.
Data Source
AI summary
Aspects of this disclosure provide a device, system, and method for analyzing text. In an embodiment, a system is configured to convert characters of the text into a numerical time series signal. The numerical time series signal includes a time series conversion of the characters in numerical format. The system is further configured to generate a waveform with extracted information from the numerical time series signal. The extracted information having features based on politeness in language, a quantifiable use of punctuations, a quantifiable use of conjunctions, use of idioms, or a combination thereof. The system is additionally configured to determine whether the text is written by a specific user based on an analysis of the waveform against a threshold.