Message Suitability Classification via ML Token Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing messaging technologies lack the ability to automatically identify and prevent the sending of messages that may be unsuitable due to offensive language, confidential information disclosure, or other sensitive topics.
Innovation Solution
A system and method that utilize machine learning classification models to analyze tokens extracted from messages before transmission, determining the intended recipients and selecting appropriate models trained on prior messages to classify the message suitability, generating an alert or blocking the message if it is deemed unsuitable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning classification models are used to automatically analyze messages before transmission, then message suitability is improved, but system complexity increases
Solution Approach 1:
The system performs message analysis before transmission by extracting tokens from the message, determining intended recipients, selecting appropriate machine learning models, and classifying message suitability in advance. This preliminary action prevents unsuitable messages from being sent, improving reliability while managing complexity through automated pre-processing.
Solution Approach 2:
The patent introduces an intermediary classification system between the user and the message transmission. This intermediary automatically analyzes message content using machine learning models, determining suitability based on extracted tokens and recipient information, thereby improving message reliability without requiring direct human intervention for each message.
2Object-generated harmful factors
If automated message analysis is implemented, then harmful factors are reduced, but processing time increases
Solution Approach 1:
The system extracts tokens from messages and determines recipient information before full transmission, performing preliminary analysis to identify potential harmful content. By conducting this analysis in advance using automated machine learning classification, the system reduces harmful factors while minimizing time loss through efficient pre-processing rather than post-transmission review.
Solution Approach 2:
The patent replaces manual message review with automated machine learning classification systems. The computer hardware automatically extracts tokens, selects appropriate models, and classifies message suitability, substituting mechanical human analysis with automated electronic processing that reduces both harmful factor detection time and overall processing delay.
3Measurement precision
If machine learning models are trained on prior messages, then classification accuracy is improved, but data processing requirements increase
Solution Approach 1:
The system trains machine learning classification models in advance using tokens extracted from prior messages and combined documents. This preliminary training action improves classification accuracy for future message analysis, allowing the system to make more precise suitability determinations while the training data processing occurs separately from real-time message transmission.
Solution Approach 2:
The patent segments the data processing into distinct phases: training data collection from prior messages, model training using combined documents, and real-time classification of new messages. This segmentation allows comprehensive data processing for training to occur separately, improving classification accuracy without burdening the real-time message transmission system with excessive data processing requirements.
Data Source
AI summary
Cognitive determination of whether a message is suitable for sending over a data communications network can include extracting tokens from the message prior to transmitting the message. One or more intended recipients of the message can be determined from the tokens. A machine learning classification model corresponding to the one or more recipients of the message can be selected. The machine learning classification model can be constructed based on tokens extracted from prior messages, which are combined to create a plurality of documents for training the machine learning classification model. The one or more tokens extracted from the message can be classified using the machine learning classification model. An alert message can be generated in response to determining based on the classifying that the message is unsuited for sending.