Model Version Upgrades Using Dynamic Schema Evolution

Solution Overview

Problem

In data science applications like fraud analytics and compliance checking, frequent changes in data and model outputs require substantial manual effort to maintain compatibility with downstream applications, leading to time-consuming schema updates.

Innovation Solution

A zero-code approach utilizing a configuration-driven dynamic schema evolution (CDDSE) that dynamically merges different data types and generates schemas based on user-defined configurations, reducing the need for manual schema editing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual schema updates are performed to maintain compatibility with downstream applications when data and model outputs change frequently, then compatibility is maintained, but time consumption and maintenance effort increase substantially

Engineering Contradiction:

ImprovecompatibilityVSAvoidtime consumption

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent implements dynamic schema evolution that automatically adapts schemas when data or model outputs change. The system monitors changes in data types and model outputs, and dynamically updates schemas without requiring manual intervention, thereby maintaining compatibility while reducing time consumption associated with frequent manual updates.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs self-updating of schemas through automated change detection and schema evolution mechanisms. When changes are detected in source data or model outputs, the system automatically generates and applies updated schemas, eliminating the need for manual schema maintenance and reducing the time and effort required to maintain compatibility.

Inventive Principle:

Principle #25Self-service

2Manufacturing precision

If manual schema editing is performed frequently to accommodate data changes, then schema accuracy is maintained, but maintenance costs and operational complexity increase

Engineering Contradiction:

Improveschema accuracyVSAvoidoperational complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent incorporates validation mechanisms that check schema changes before they are applied, ensuring accuracy is maintained. The system validates proposed schema changes against downstream application requirements and data constraints, preventing inaccurate schemas from being deployed while reducing the complexity of manual validation procedures.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements feedback loops that monitor the impact of schema changes on downstream applications. When schema changes are made, the system validates their correctness and provides feedback for correction if issues are detected, thereby maintaining schema accuracy while reducing the operational complexity through automated validation rather than manual checking.

Inventive Principle:

Principle #23Feedback

3Adaptability or versatility

If traditional model version upgrading approaches are used, then model updates can be implemented, but substantial manual effort and downtime are required

Engineering Contradiction:

Improvemodel update capabilityVSAvoidmanual effort

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent implements dynamic schema evolution that automatically adapts schemas when data or model outputs change. The system monitors changes in data types and model outputs, and dynamically updates schemas without requiring manual intervention, thereby maintaining compatibility while reducing time consumption associated with frequent manual updates.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs self-updating of schemas through automated change detection and schema evolution mechanisms. When changes are detected in source data or model outputs, the system automatically generates and applies updated schemas, eliminating the need for manual schema maintenance and reducing the time and effort required to maintain compatibility.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS12579124B2Zero-code approach for model version upgrades

Publication Date: 2026.03.17 HONEYWELL INTERNATIONAL INC

AI summary

Systems and methods for a zero-code approach for model version upgrades are described herein. For example, a system includes memory. The system further includes a processor that receives source data from data sources; applies an ingestion process to the source data for storing storage data in a data storage repository; generates models from the storage data; applies an output process to generate output schemas based on the generated models; and applies the output schemas for providing data. Also, the source data and the models changes over time leading to different data type collections. Further, at least one of applying the ingestion and the output process includes performing a dynamic merge on the different data type collections to create a merged data type collection; acquiring a user-defined configuration for the merged data type collection; and generating an output schema to the merged data type collection based on the user-defined configuration.