Machine Learning Ranking for Hybrid UI Workflow Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing rules-based recommendation engines for user interfaces are inefficient and unreliable, leading to wasted computing resources due to the presentation of irrelevant results, as they lack personalization and adaptability based on user behavior.

Innovation Solution

A hybrid approach combining machine learning and rules-based logic to generate and rank recommendations, utilizing a machine learning model to enhance the accuracy and efficiency of user interface workflows by personalizing recommendations based on user behavior and domain features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If rules-based recommendation engines are used to provide static recommendation results in a list form, then the system structure is simple and easy to implement, but the recommendations lack personalization and computing resources are wasted presenting irrelevant results

Engineering Contradiction:

Improvesystem implementation simplicityVSAvoidrecommendation personalization accuracy

Core Design Contradiction:

Ease of manufactureVSReliability

Solution Approach 1:

The patent merges rules-based recommendation logic with machine learning-based personalization in a hybrid system. The rules-based component maintains structural simplicity and provides foundational recommendations, while the machine learning component adds personalization and relevance ranking. This combination resolves the contradiction by integrating the simplicity of rules-based systems with the accuracy of ML-based personalization.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The recommendation system is segmented into multiple components: a rules-based recommendation engine for generating initial recommendations, a machine learning model for personalization and ranking, and a presentation layer that filters and displays relevant results. This segmentation allows each component to optimize for its specific function while working together to resolve the overall contradiction.

Inventive Principle:

Principle #1Segmentation

2Ease of operation

If rules-based recommendation engines present static recommendation results, then the system is easy to operate, but users must interact with multiple irrelevant results before finding sufficient recommendations

Engineering Contradiction:

Improveuser interface simplicityVSAvoiduser interaction time

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The machine learning model performs preliminary action by pre-ranking and filtering recommendation results based on user preferences and behavior patterns before they are presented to the user. This preliminary personalization and sorting reduces the number of irrelevant results users must interact with, saving time while maintaining ease of operation through a simplified interface.

Inventive Principle:

Principle #10Preliminary action

3Reliability

If machine learning models are added to personalize recommendations, then recommendation accuracy improves, but system complexity increases

Engineering Contradiction:

Improverecommendation accuracyVSAvoidsystem architecture complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system merges rules-based and machine learning-based approaches in a coordinated architecture where the rules-based component generates initial recommendations and the machine learning component personalizes and ranks them. This integration achieves high recommendation accuracy while managing complexity through a structured hybrid architecture rather than using ML alone.

Inventive Principle:

Principle #5Merging (Combining)

4Productivity

If irrelevant recommendation results are presented via user interface, then the system processes all recommendations, but computing resources are wasted

Engineering Contradiction:

Improverecommendation processing throughputVSAvoidcomputing resource consumption

Core Design Contradiction:

ProductivityVSLoss of energy

Solution Approach 1:

The machine learning model performs preliminary filtering and ranking of recommendation results before they are presented to the user. This preliminary action identifies and prioritizes relevant recommendations, reducing the number of irrelevant results that need to be processed and displayed, thereby saving computing resources while maintaining processing throughput.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system extracts and removes irrelevant recommendation results from the final presentation set using machine learning-based filtering. By taking out irrelevant results before presentation, the system reduces computing resource consumption on displaying and processing unnecessary recommendations while maintaining productivity in generating and filtering relevant ones.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS20250238705A1Machine learning and rules-based recommendations for user interface workflows

Publication Date: 2025.07.24 OPTUM INC

AI summary

Various embodiments of the present disclosure provide machine learning and rules-based recommendations for user interface workflows. In one example, an embodiment provides for generating a set of recommendation data objects for a user identifier associated with a user interface based on a set of predefined rules associated with input data provided via a user interface workflow associated with the user interface, generating a ranked version of the set of recommendation data objects using a machine learning model, and initiating a rendering of a set of selectable graphical elements via the user interface based on the ranked version of the set of recommendation data objects.