Image Processing Model Versioning and Hyperparameter Mutation Tracking

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional neural-network based machine-learning models face challenges in interpretability, require extensive time and computational resources for construction and optimization, and have cumbersome storage and management needs due to their complex architecture.

Innovation Solution

A system and method for managing image processing models that includes a model abstraction layer for selection, mutation, and construction, versioning of neural-network based models, and a parameter archival storage system for read-optimized compression storage, allowing for reduced storage and efficient mutation tracking of hyperparameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If machine-learning neural-network based models are used for complex image processing functions, then the processing capability and accuracy are improved, but the storage requirements and management complexity increase

Engineering Contradiction:

Improveprocessing capabilityVSAvoidmanagement complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the model management system into distinct functional modules: version control system, hyperparameter tracking system, and compressed storage system. Each module handles specific aspects of model management independently, reducing overall system complexity while maintaining comprehensive control over neural network models

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary layer between the neural network models and the management system, using standardized data structures and interfaces to mediate interactions. This intermediary abstraction simplifies management operations while preserving the complex processing capabilities of the underlying models

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If conventional neural-network models are constructed and optimized, then the model performance is improved, but the time and computational resources required increase

Engineering Contradiction:

Improvemodel performanceVSAvoidconstruction time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent implements preliminary version control and hyperparameter configuration systems that prepare model structures and parameters before actual training. This preliminary setup organizes the construction process, reducing the time and computational resources needed during the actual model optimization phase

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent incorporates feedback mechanisms through hyperparameter tracking and version comparison systems. These systems provide continuous feedback during model construction and optimization, enabling efficient adjustments that reduce computational resources and time while maintaining or improving model performance

Inventive Principle:

Principle #23Feedback

3Loss of information

If neural-network models are stored with full architecture details, then the model information completeness is improved, but the storage space required increases

Engineering Contradiction:

Improvemodel information completenessVSAvoidstorage space

Core Design Contradiction:

Loss of informationVSQuantity of substance

Solution Approach 1:

The patent extracts and separates critical model information (architecture definitions, hyperparameter configurations, version metadata) from the full model data. This extracted information is stored in optimized formats that maintain completeness of essential model information while significantly reducing storage space requirements

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent transforms model storage from raw data formats to compressed parameter representations. By changing the storage parameters and using efficient data structures for hyperparameters and architecture definitions, the system maintains full model information completeness while achieving significant storage space reduction

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11481633B2Electronic system for management of image processing models

Publication Date: 2022.10.25 BANK OF AMERICA CORP

AI summary

Embodiments of the invention are directed to systems, methods, and computer program products for an electronic system for management of image processing model database. The system is configured for versioning machine-learning neural-network based image processing models and identifying and tracking mutations in hyper parameters amongst versions of image processing models. The system is configured to determine that a second image processing model is a version of a first image processing model. The system is further configured to map the mutations in hyper parameters between the first plurality of hyper parameters of the first image processing model and the second plurality of hyper parameters associated with the second image processing model.