News‎ > ‎

Synaptic Versus Neuronal Stochasticity explained

posted Nov 3, 2016, 7:05 PM by Khaled salama   [ updated Nov 3, 2016, 7:18 PM ]
In yet another very interesting work, PhD student Rawan Naous explores stochasticity in the cortex based on either the synaptic or neuronal components. The hardware emulation of these stochastic neural networks are currently being extensively studied using resistive memories or memristors.  Aside from the size and area perspective, the impact on the system performance, in terms of accuracy, recognition rates, and learning, among these two approaches and where the memristor would fall into place are the main comparison points considered.  For more details check:

Rawan Naous, Maruan AlShedivat, Emre Neftci, Gert Cauwenberghs and Khaled Nabil Salama, " Memristor-based neural networks: Synaptic versus neuronal stochasticity"AIP Advances 6, 111304 (2016); http://dx.doi.org/10.1063/1.4967352

This expands our earlier work:
  1. Rawan Naous, Maruan Al-Shedivat, and Khaled Nabil Salama, Stochasticity Modeling in MemristorsIEEE Transactions on Nanotechnology (TNANO), vol. 15, no. 1, pp. 15-28, 2016 DOI:10.1109/TNANO.2015.2493960
  2.  Rawan Naous., Maruan Al-Shedivat, Emri Neftci, Gert Cauwenberghs and Khaled Nabil Salama "Stochastic Synaptic Plasticity with Memristor Crossbar Arrays", IEEE Int'l Symposium on Circuits & Systems (ISCAS), to be held in Montreal, Canada, 2016
  3. Maruan Al-Shedivat, Rawan Naous,Gert Cauwenberghs, and Khaled Nabil Salama, Memristors Empower Spiking Neurons with Stochasticity, IEEE Journal of Emerging technologies in circuits and systems, VOL. 5, NO. 2, 242-253, JUNE 2015
  4. Maruan Al-Shedivat, Rawan Naous., Emre Neftciy, Gert Cauwenberghsy and Khaled N. Salama., "Inherently Stochastic Spiking Neurons for Probabilistic Neural Computation', 7th International IEEE EMBS Neural Engineering Conference (NER'15), Montpellier, France on April 22-24, 2015.

Congrats Rawan!
Comments