On-device intelligence using a low-power deep learning method- WIE ILC 2021

3 views
Download
  • Share

Speaker: Foroozan Karimzadeh, Graduate Research Assistant, Georgia Institute of Technology

The increase in the number of edge devices such as mobile, wearable and Internet of Things (IoT) has led to the emergence of edge computing where the computations are performed on the device. In recent years, deep neural networks (DNNs) have become the state-of-the-art methods in a broad range of applications, from image recognition, to health care data analysis and self-driving cars. However, neural network models are typically large and computationally expensive and therefore not deployable on power and memory constrained edge devices. In this presentation, I'll talk about novel hardware-friendly approaches towards sparse and quantized neural networks. Using these method a significant amount of power savings can be achieved which enables DNNs to be deployable on mobile applications. As an example, we can use the proposed low power DNN for home monitoring devices for patients to make them cheaper and more accurate.

Speaker: Foroozan Karimzadeh, Graduate Research Assistant, Georgia Institute of Technology

The increase in the number of edge devices such as mobile, wearable and Internet of Things (IoT) has led to the emergence of edge computing where the computations are performed on the device. In recent years, deep neural networks (DNNs) have become the state-of-the-art methods in a broad range of applications, from image recognition, to health care data analysis and self-driving cars...

Advertisment

Advertisment