ML System for Multi-Modal UI Data Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing analytics models are inadequate in interpreting multi-dimensional data from user interfaces, struggling to accurately decipher and make predictions from varied information sources in computing environments.
Innovation Solution
A computer-implemented system utilizing a machine learning engine that processes data records with categorical and free-form textual data, employing natural language processing and graph convolutional networks to derive categorical and continuous attributes, generate entity graphs, and provide ensemble model predictions for intelligent query responses and recommended digital actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing analytics models are used to process user interface data, then the system structure is simple, but the models struggle to accurately decipher multi-dimensional data and make predictions
Solution Approach 1:
The patent combines multiple machine learning models (natural language processing model, graph convolutional network, and ensemble model) into an integrated system. The NLP model processes free-form text data, the GCN processes categorical variables and entity relationships, and their outputs are combined through an ensemble model to generate predictions. This merging of multiple specialized models resolves the contradiction by achieving high prediction accuracy through comprehensive multi-dimensional data processing while maintaining a structured modular architecture.
Solution Approach 2:
The patent segments the data processing task into distinct components: free-form text data is processed separately by the NLP model, while categorical variables and entity relationships are processed separately by the GCN. This segmentation allows each model to specialize in handling specific data types and dimensions, improving overall prediction accuracy without creating a monolithic complex system.
2Adaptability or versatility
If multiple types of input data are processed to improve understanding of multi-dimensional data, then the predictive performance improves, but the system complexity increases
Solution Approach 1:
The patent creates a universal machine learning system that can handle multiple data types (free-form text, categorical variables, entity relationships) through a unified architecture. The ensemble model serves as a universal predictor that integrates outputs from different specialized models, allowing the system to adapt to various data dimensions and input types without requiring separate independent systems for each data type.
Solution Approach 2:
The ensemble model acts as an intermediary that bridges the NLP model and GCN model. It receives processed outputs from both models and combines them to generate final predictions. This intermediary component enables the system to effectively integrate information from multiple data sources and processing pathways, enhancing versatility while maintaining a manageable system structure through clear separation of concerns.
3Measurement precision
If natural language processing and graph convolutional networks are used to process different data types, then the predictive accuracy improves, but the computational requirements and processing time increase
Solution Approach 1:
The patent segments the computational workload by assigning different data processing tasks to specialized models that can operate in parallel. The NLP model processes free-form text data independently while the GCN processes categorical variables and entity relationships simultaneously. This segmentation of computational tasks allows for more efficient resource utilization and can reduce overall processing time compared to a sequential single-model approach, while maintaining high prediction accuracy through the ensemble integration of both model outputs.
Data Source
AI summary
In one or more aspects, there is provided a machine learning system and method for generating recommended electronic actions on user interfaces of requesting user interface query devices. In one or more aspects there is provided a machine learning based engine and device to process multiple modes of input user interface data utilizing natural language processing and machine learning models for processing different modes and determining intelligent computerized responses and digital actions based on the machine learning processing.