Neural Network Hardware Accelerator at the Edge
In the recent times, there has been an increase in the number of applications ranging from robotics, speech processing, IoT-based devices and so on using neural network computations. This has led to an increased demand for specialized hardware architectures targetting neural network infernce and training computation from the server to the extreme edge level.
This project is targetted to work on designing novel hardware accelerators for application targeting wearables. The idea is to design accelerators having extremely low area and power output.
One of the previous works is the Shenjing accelerator which is an SNN-based accelerator targeting wearable applications.
One of the recent works published is the REACT hardware accelerator, which is a reconfigurable hardware accelerator which allows heterogeneous cores supporting both training and inference.
