Machine Learning Scoring for Adaptive Visualization Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional visualization recommendation tools in content creation software are inefficient and labor-intensive, often relying on manually designed heuristics that fail to adapt to diverse dataset features and user needs, leading to suboptimal visualization configurations.

Innovation Solution

A visualization recommendation system that utilizes machine learning techniques to generate recommendation scores for visualization configurations by analyzing meta-features of datasets and configurations, employing wide and deep scoring models to automate the selection of suitable visualization tools based on dataset characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If manually designed heuristics are used for visualization recommendation, then the system is easier to implement and maintain, but the system cannot evaluate a wide variety of visualization configurations and becomes rapidly outdated

Engineering Contradiction:

Improveease of implementationVSAvoidability to evaluate visualization configurations

Core Design Contradiction:

Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent replaces manual heuristic design with automated machine learning models (wide and deep scoring models) that automatically evaluate visualization configurations. The system extracts meta-features from datasets and configurations, then uses neural networks to score and rank visualizations, eliminating the need for hand-crafted rules while enabling comprehensive evaluation of any configuration.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system transforms manual heuristic parameters into automated machine learning parameters. Instead of coding specific rules, the system learns patterns from data through training on visualization effectiveness. The scoring models automatically adapt to new visualization types and dataset characteristics without requiring manual reconfiguration of heuristics.

Inventive Principle:

Principle #35Parameter changes

2Device complexity

If manually designed heuristics are used, then development effort is reduced initially, but additional heuristics require extensive development and testing effort

Engineering Contradiction:

Improvedevelopment effortVSAvoidease of adding new heuristics

Core Design Contradiction:

Device complexityVSEase of manufacture

Solution Approach 1:

The machine learning system performs self-learning and self-updating without requiring manual intervention to add new heuristics. The deep and wide scoring models automatically adapt to new visualization configurations and dataset types through training on additional data, eliminating the need for technicians to develop and test new heuristic rules for each improvement.

Inventive Principle:

Principle #25Self-service

3Device complexity

If manually designed tools are used, then the system structure is simpler, but the tools become rapidly outdated requiring additional time and effort to update

Engineering Contradiction:

Improvesystem structureVSAvoidrelevance over time

Core Design Contradiction:

Device complexityVSDuration of action of stationary object

Solution Approach 1:

The system transitions from static manual heuristics to dynamic machine learning models that continuously learn and adapt. The scoring models can be retrained on new data to incorporate emerging visualization best practices and adapt to changing data characteristics, maintaining relevance over time without requiring complete system redesign.

Inventive Principle:

Principle #15Dynamics

4Device complexity

If users manually analyze dataset features to select visualizations, then the system requires fewer automated components, but users spend a long amount of time analyzing features

Engineering Contradiction:

Improveautomation componentsVSAvoidtime for feature analysis

Core Design Contradiction:

Device complexityVSLoss of time

Solution Approach 1:

The system performs preliminary feature extraction and visualization scoring automatically before the user needs to make a selection. The machine learning models pre-evaluate multiple visualization configurations against the dataset characteristics, ranking them in advance so the user can directly access recommended visualizations without manually analyzing features.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12462169B2Machine learning techniques for generating visualization recommendations

Publication Date: 2025.11.04 ADOBE INC

AI summary

A visualization recommendation system generates recommendation scores for multiple visualizations that combine data attributes of a dataset with visualization configurations. The visualization recommendation system maps meta-features of the dataset to a meta-feature space and configuration attributes of the visualization configurations to a configuration space. The visualization recommendation system generates meta-feature vectors that describe the mapped meta-features, and generates configuration attribute sets that describe the attributes of the visualization configurations. The visualization recommendation system applies multiple scoring models to the meta-feature vectors and configuration attribute sets, including a wide scoring model and a deep scoring model. In some cases, the visualization recommendation system trains the multiple scoring models using the meta-feature vectors and configuration attribute sets.