Model Catalog Ontology for Automated ML Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Identifying an appropriate machine learning model to derive insights from a particular dataset can be difficult and resource-intensive, especially when multiple models of varying types are stored in a model store.

Innovation Solution

A cloud infrastructure system implements a model catalog with a domain ontology store that recommends models based on dataset features. The system obtains and parses the dataset, compares it with terms associated with multiple models, identifies recommended models, and executes simulations to provide insights.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple machine learning models are stored in a model store for diverse data processing needs, then the system's adaptability and versatility improve, but the complexity of selecting an appropriate model increases

Engineering Contradiction:

Improvemodel selection capabilityVSAvoidmodel selection process

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system automatically parses dataset features, compares them with model metadata, and generates model recommendations without requiring manual intervention. The model store self-services by maintaining structured metadata that enables automated matching, reducing the complexity of model selection for clients.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

Metadata acts as an intermediary between datasets and machine learning models. The metadata contains structured information about model inputs, outputs, and characteristics, enabling automated comparison and recommendation without direct complex interactions between clients and multiple models.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If a client manually reviews and executes multiple potential models to identify an appropriate model, then the reliability of model selection improves, but the time and resources required increase significantly

Engineering Contradiction:

Improvemodel selection accuracyVSAvoidmodel review time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

Metadata is prepared in advance for each model, containing structured information about inputs, outputs, and model characteristics. This preliminary organization enables rapid automated comparison when a client submits a dataset, eliminating the need for time-consuming manual model review while maintaining selection accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system provides automated model recommendations with confidence scores based on metadata matching. Clients can review these pre-filtered recommendations, which are generated through systematic comparison of dataset features with model metadata, reducing both time and resources while maintaining reliable model selection.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12299005B2Model mining and recommendation engine with simulation interfaces

Publication Date: 2025.05.13 ORACLE INT CORP

AI summary

The present embodiments relate to data processing model recommendation and review of a portion of data using a recommended model. A model catalog executing on a cloud infrastructure (CI) system can parse data from an obtained dataset identifying aspects of the dataset. The parsed data from the dataset can be compared with a plurality of potential models stored in a domain ontology store of the model catalog to identify one or more recommended models. Review output data can be generated using the dataset and any of the recommended models. The review output data resulting from the recommended model can be provided to the client for the client to either accept or reject the model.