Edge ML Model Optimization for Building Management Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing building management systems (BMS) face challenges in executing machine-learning models on edge devices due to limited processing capabilities and hardware constraints.
Innovation Solution
A BMS that includes processors and non-transitory memory, capable of receiving optimization criteria to transform machine-learning model parameters, generating a second machine-learning model suitable for a target platform, and retraining the model based on accuracy thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine-learning models with high computational requirements are executed on edge devices, then model accuracy can be maintained, but the limited processing capabilities and hardware constraints of edge devices cannot support such models
Solution Approach 1:
The patent transforms data types of model parameters from high-precision formats (e.g., 32-bit or 64-bit floating point) to lower-precision formats (e.g., 8-bit, 16-bit, or 32-bit floating point). This parameter change reduces the computational burden and memory requirements, enabling edge devices with limited processing capabilities to execute machine-learning models while maintaining acceptable accuracy levels through subsequent retraining processes
2Device complexity
If model parameters are transformed to lower precision data types to fit edge device constraints, then the model can be executed on resource-constrained devices, but model accuracy deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the transformed model is evaluated using a verification dataset to measure accuracy. When accuracy falls below a predetermined threshold, the system automatically triggers retraining using a training dataset. This closed-loop feedback process continues iteratively, adjusting model parameters and retraining until the model achieves satisfactory accuracy on edge devices with transformed data types
Solution Approach 2:
The patent performs preliminary transformation of data types before deploying the model to edge devices. By proactively converting parameters to lower-precision formats and pre-evaluating the transformed model's accuracy, the system prepares the model in advance for edge device execution, identifying necessary retraining requirements before deployment and ensuring the model is optimized for the target hardware constraints
Data Source
AI summary
Systems and methods described herein are directed to the optimization and implementation of machine-learning models on edge devices. A building management system can receive optimization criteria to optimize a first machine-learning model for a target platform. The first machine-learning model has one or more model parameters. The building management system transforms, based on the target platform and the optimization criteria, at least one datatype of the one or more model parameters of the first machine-learning model to generate a second machine-learning model. The building management system determines, using a verification dataset, an accuracy of the second machine-learning model, and retrains the second machine-learning model using a training dataset responsive to the accuracy being less than a predetermined threshold.