Neuromorphic Deep Learning Hardware Accelerator
Lead: Institute of Neuroinformatics, University of Zurich
This work focuses on the hardware implementation of artificial intelligent deep networks to create stimulation signals for the stimulation platform. The work will be based on the hardware deep network accelerator development in the group and will include the refinement of these deep network algorithms so that the training is hardware-aware and the architectures match the hardware constraints of the eventual low-power embedded platform.
The output of this platform will be interfaced to the recording and stimulation electronics. Real-time performance will be key for the eventual closed-loop operation of the entire system.