Neural Network Application Navigation Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for guiding users through applications are inefficient and inaccurate, as they often reflect developer assumptions rather than actual user interactions, and become obsolete with application updates or changes in usage patterns.

Innovation Solution

A system using linked neural networks to predict the next page in an application by dividing the prediction process into smaller sub-problems, with a state predictor and a data predictor, trained with actual usage data to provide updated and context-aware recommendations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If training materials are used to guide users through an application, then users can learn how to interact with the application, but the training materials become outdated when the application is updated or usage patterns change

Engineering Contradiction:
Improveaccuracy of navigation guidanceVSAvoidability to adapt to application updates and changing usage patterns
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system transitions from static training materials to dynamic neural network models that continuously adapt to changing usage patterns. The neural networks are retrained on new interaction data, enabling the navigation guidance to evolve automatically with application updates and changing user behaviors, thus resolving the contradiction between reliability and adaptability

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system implements feedback loops where user interactions are collected and used to retrain the neural networks. This continuous feedback mechanism ensures that the navigation recommendations remain accurate and up-to-date, automatically adapting to application changes and new usage patterns without requiring manual updates to training materials

Inventive Principle:
Principle #23Feedback

2Measurement precision

If multiple neural networks are used to predict next page, then prediction accuracy improves, but system complexity increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The prediction task is segmented into multiple specialized neural networks, each handling specific aspects of navigation prediction. This segmentation allows each network to focus on particular patterns, improving overall prediction accuracy while organizing complexity into manageable, modular components that can be independently trained and maintained

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3762821B1Neural network systems and methods for application navigation
Publication Date: 2024.06.05 AMADEUS SAS
  • EP3762821B1 patent drawingFigure 1
  • EP3762821B1 patent drawingFigure 2A
  • EP3762821B1 patent drawingFigure 2B

AI summary

The present disclosure relates to systems and methods for application navigation using neural networks. The disclosed systems and methods can perform operations including providing an application comprising application states that operate on data objects to generate pages, determining a current application state and current data object, predicting a next page using an application state-predicting neural network and a data object-predicting neural network and providing instructions to display an indication of the predicted next page. Predicting the next page can include predicting a next application state using the application state-predicting neural network and a first input vector, and predicting a next data object using the data object-predicting neural network, a second input vector, and the next application state.