AI Model Management Platform for Dynamic Multi-Model Replacement

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Managing multiple machine learning models across various features in a client platform is complex due to varying relationships and resource demands, making it difficult to optimize performance and resource usage efficiently.

Innovation Solution

A model management system that automates performance measurement and dynamically replaces models based on the service provision environment, considering device information and usage patterns to optimize resource allocation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple AI models are managed for various features in a client platform, then feature functionality and adaptability are improved, but device complexity and resource consumption increase

Engineering Contradiction:

Improvefeature functionalityVSAvoidmodel management complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

A model management platform is introduced as an intermediary system between the client device and multiple AI models. This platform automatically manages model downloads, updates, deletions, and performance measurements, shielding users from the complexity of managing multiple models while enabling diverse feature functionality through coordinated model orchestration.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If multiple AI models are downloaded and stored for different features, then service functionality is improved, but memory usage and storage requirements increase

Engineering Contradiction:

Improveservice functionalityVSAvoidmemory usage

Core Design Contradiction:

Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The system dynamically manages AI model storage by automatically downloading models when needed for specific features and deleting them when no longer required. The model management platform monitors feature usage and model performance metrics, adjusting the set of stored models adaptively to balance functionality with memory constraints, ensuring models are present only when their corresponding features are active or scheduled.

Inventive Principle:

Principle #15Dynamics

3Measurement precision

If model performance is measured for each AI model, then model selection accuracy is improved, but measurement time and processing overhead increase

Engineering Contradiction:

Improvemodel performance measurementVSAvoidmeasurement time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The model management platform performs preliminary performance measurements of AI models during off-peak periods or in the background, before models are actually needed for feature execution. By pre-measuring performance metrics such as accuracy, speed, and resource consumption, the system has measurement data ready when model selection decisions must be made, eliminating measurement delays during critical service delivery moments.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260030555A1Methods, computer devices, and non-transitory computer readable media for managing models and dynamic replacement of multiple models

Publication Date: 2026.01.29 LINE PLUS

AI summary

Disclosed is a model management method executed by a computer device, the computer device including at least one processor configured to execute computer-readable instructions included in a memory, and the model management method including integrally managing, by the at least one processor, a plurality of Artificial Intelligence (AI) models through a platform of a client, each respective AI model among the plurality of AI models being related to a corresponding feature among a plurality of features included in an application installed at the client.