18223888. METHOD AND SYSTEM FOR OPTIMIZING NEURAL NETWORKS (NN) FOR ON-DEVICE DEPLOYMENT IN AN ELECTRONIC DEVICE simplified abstract (Samsung Electronics Co., Ltd.)
METHOD AND SYSTEM FOR OPTIMIZING NEURAL NETWORKS (NN) FOR ON-DEVICE DEPLOYMENT IN AN ELECTRONIC DEVICE
Organization Name
Inventor(s)
Ashutosh Pavagada Visweswara of Bengaluru (IN)
Arun Abraham of Bengaluru (IN)
Vikram Nelvoy Rajendiran of Bengaluru (IN)
Rajath Elias Soans of Bengaluru (IN)
METHOD AND SYSTEM FOR OPTIMIZING NEURAL NETWORKS (NN) FOR ON-DEVICE DEPLOYMENT IN AN ELECTRONIC DEVICE - A simplified explanation of the abstract
This abstract first appeared for US patent application 18223888 titled 'METHOD AND SYSTEM FOR OPTIMIZING NEURAL NETWORKS (NN) FOR ON-DEVICE DEPLOYMENT IN AN ELECTRONIC DEVICE
Simplified Explanation
The patent application describes systems and methods for optimizing neural networks for on-device deployment in an electronic device. Here is a simplified explanation of the abstract:
- The method involves receiving multiple neural network (NN) models.
- At least two NN models are fused together based on at least one layer from each model, resulting in a fused NN model.
- The fused NN model is analyzed to identify any redundant layers.
- The identified redundant layers are removed from the fused NN model, resulting in an optimized NN model.
Potential applications of this technology:
- On-device deployment of neural networks in electronic devices.
- Optimizing the performance and efficiency of neural networks for specific tasks.
Problems solved by this technology:
- Reducing the size and complexity of neural networks for on-device deployment.
- Improving the efficiency and speed of neural network computations on electronic devices.
Benefits of this technology:
- Enables efficient deployment of neural networks on electronic devices without relying on cloud-based processing.
- Reduces the memory and computational requirements of neural networks, making them more suitable for resource-constrained devices.
- Improves the overall performance and responsiveness of neural network applications on electronic devices.
Original Abstract Submitted
Provided are systems and methods for optimizing neural networks for on-device deployment in an electronic device. A method for optimizing neural networks for on-device deployment in an electronic device includes receiving a plurality of neural network (NN) models, fusing at least two NN models from the plurality of NN models based on at least one layer of each of the at least two NN models, to generate a fused NN model, identifying at least one redundant layer from the fused NN model, and removing the at least one redundant layer to generate an optimized NN model.
