Dynamic Time-Based Custom Model Generation

Solution Overview

Problem

The challenge is to efficiently generate and train machine learning models within predetermined time constraints and resource limitations, as existing methods often result in models that are not fully optimized due to time and resource intensive hyperparameter tuning and training processes.

Innovation Solution

The proposed solution involves dynamic time-based custom model generation techniques, where a client provides a training dataset and constraints, and a computing device optimizes model generation and training by determining a modified training dataset and hyperparameter values, allowing for training within the specified time frame.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If comprehensive hyperparameter tuning and extensive training data are used to improve model accuracy, then model accuracy is improved, but training time and resource consumption increase exponentially

Engineering Contradiction:

Improvemodel accuracyVSAvoidtraining time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-processing the training dataset to determine a modified training dataset and pre-determining hyperparameter values before actual model training begins. This preparation work is done in advance to enable the model training to proceed efficiently within the constrained time period, resolving the contradiction between achieving high accuracy and limiting training time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system extracts and utilizes existing hyperparameter values from the training dataset or pre-defined configurations, rather than performing exhaustive hyperparameter tuning. This extraction approach eliminates the time-consuming hyperparameter optimization process while still providing sufficient model accuracy, thereby resolving the contradiction between model quality and training duration.

Inventive Principle:

Principle #2Taking out (Extraction)

2Manufacturing precision

If comprehensive hyperparameter tuning and extensive training data are used to improve model accuracy, then model accuracy is improved, but resource consumption increases exponentially

Engineering Contradiction:

Improvemodel accuracyVSAvoidresource consumption

Core Design Contradiction:

Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary actions by pre-processing the training dataset to determine a modified training dataset and pre-determining hyperparameter values before actual model training begins. This preparation work is done in advance to enable the model training to proceed efficiently within the constrained time period, resolving the contradiction between achieving high accuracy and limiting training time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system extracts and utilizes existing hyperparameter values from the training dataset or pre-defined configurations, rather than performing exhaustive hyperparameter tuning. This extraction approach eliminates the time-consuming hyperparameter optimization process while still providing sufficient model accuracy, thereby resolving the contradiction between model quality and training duration.

Inventive Principle:

Principle #2Taking out (Extraction)

3Productivity

If training is stopped after a predetermined time period to meet client deadlines, then time constraints are satisfied, but model accuracy deteriorates due to insufficient training

Engineering Contradiction:

Improvemodel generation speedVSAvoidmodel accuracy

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The system performs preliminary actions by pre-processing the training dataset to determine a modified training dataset and pre-determining hyperparameter values before actual model training begins. This preparation work is done in advance to enable the model training to proceed efficiently within the constrained time period, resolving the contradiction between achieving high accuracy and limiting training time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts the training process by using a modified training dataset and pre-determined hyperparameter values that are optimized for the specific time constraint. This dynamic adaptation allows the model to achieve sufficient accuracy within the predetermined time period, resolving the contradiction between meeting deadlines and achieving model quality.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20250157210A1Techniques for dynamic time-based custom model generation

Publication Date: 2025.05.15 ORACLE INT CORP

AI summary

Techniques are disclosed for dynamic time-based custom model generation as part of infrastructure-as-a-service (IaaS) environment. A custom model generation service may receive a set of training data and a time-based constraints for training a machine learning model. The custom model generation service may subsample the training data and generate a set of optimized tuned hyperparameters for a machine learning model to be trained using the subsampled training data. An experimental interval time of training is determined and the machine learning model is trained on the subsampled training data according to the optimized tuned hyperparameters over a set of training intervals similar to the experimental time interval. A customized machine learning model trained in the time-based constraint is output. The hyperparameter tuning may be performed using a modified mutating genetic algorithm for a set of hyperparameters to determine the optimized tuned hyperparameters prior to the training.