Sony group corporation (20240256856). DEPLOYING NEURAL NETWORK MODELS ON RESOURCE-CONSTRAINED DEVICES simplified abstract
Contents
DEPLOYING NEURAL NETWORK MODELS ON RESOURCE-CONSTRAINED DEVICES
Organization Name
Inventor(s)
KRISHNA PRASAD AGARA VENKATESHA Rao of BANGALORE (IN)
AKSHAY SHEKHAR Kadakol of Bangalore (IN)
PRAJOT S. Kuvalekar of Bangalore (IN)
DEV PRASAD Kode of Bangalore (IN)
DEPLOYING NEURAL NETWORK MODELS ON RESOURCE-CONSTRAINED DEVICES - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240256856 titled 'DEPLOYING NEURAL NETWORK MODELS ON RESOURCE-CONSTRAINED DEVICES
Simplified Explanation: This patent application describes a method for deploying neural network models on resource-constrained devices by partitioning the model and loading sub-models as needed for specific tasks.
Key Features and Innovation:
- Storing a model file containing a neural network model.
- Determining constraint information for deployment on electronic devices.
- Partitioning the neural network model based on constraints.
- Extracting sub-models from the neural network model.
- Loading sub-models into working memory for specific tasks.
- Applying sub-models to input data to generate results.
- Unloading sub-models from memory after use.
Potential Applications: This technology can be applied in various fields such as:
- Edge computing
- Internet of Things (IoT) devices
- Mobile applications
- Wearable technology
Problems Solved: This technology addresses the challenge of deploying complex neural network models on devices with limited resources efficiently.
Benefits:
- Improved performance on resource-constrained devices.
- Reduced memory usage.
- Faster execution of machine learning tasks.
Commercial Applications: Title: Efficient Neural Network Model Deployment for Resource-Constrained Devices This technology can be utilized in industries such as:
- Healthcare for wearable devices
- Automotive for edge computing in vehicles
- Retail for personalized customer experiences
Prior Art: Readers can explore prior art related to neural network model deployment on resource-constrained devices in academic journals, patent databases, and industry publications.
Frequently Updated Research: Stay updated on the latest advancements in neural network model optimization for resource-constrained devices to enhance performance and efficiency.
Questions about Neural Network Model Deployment on Resource-Constrained Devices: 1. How does this method improve the efficiency of deploying neural network models on resource-constrained devices? 2. What are the potential limitations of partitioning neural network models for deployment on electronic devices?
Original Abstract Submitted
a method for deploying neural network models on resource-constrained devices is provided. the method includes storing a model file that includes a neural network model and determining constraint information associated with deployment of the neural network model on the electronic device. the method further includes determining a partition of the neural network model based on the constraint information and the model file and extracting sub-models from the neural network model based on the partition. the method further includes receiving an input associated with a machine learning task and executing operations for loading a sub-model in a working memory of the electronic device, applying the sub-model on the input to generate an intermediate result, and unloading the sub-model from the working memory. the method further includes executing the operations for a next sub-model to generate an output and rendering the output. the intermediate result is an input for the next sub-model.
