Reconfigurable Neuromorphic Computing System with Memristor-Based Synapse Design

Conventional CMOS technology is slowly approaching its physical limitations and researchers are increasingly utilizing nanotechnology to both extend CMOS capabilities and to explore potential replacements. Novel memristive systems continue to attract growing attention since their reported physical realization by HP in 2008. Unique characteristics like non-volatility, re-configurability, and analog storage properties make memristors a very promising candidate for the realization of artificial neural systems. In this work, we propose a memristor-based design of bidirectional transmission excitation/inhibition synapses and implement a neuromorphic computing system based on our proposed synapse designs. The robustness of our system is also evaluated by considering the actual manufacturing variability with emphasis on process variation.

[1]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[2]  Peng Li,et al.  Nonvolatile memristor memory: Device characteristics and design implications , 2009, 2009 IEEE/ACM International Conference on Computer-Aided Design - Digest of Technical Papers.

[3]  Kanter,et al.  Associative recall of memory without errors. , 1987, Physical review. A, General physics.

[4]  Damien Querlioz,et al.  Simulation of a memristor-based spiking neural network immune to device variations , 2011, The 2011 International Joint Conference on Neural Networks.

[5]  H. C. Card,et al.  Analog CMOS deterministic Boltzmann circuits , 1993 .

[6]  Yiran Chen,et al.  Statistical memristor modeling and case study in neuromorphic computing , 2012, DAC Design Automation Conference 2012.

[7]  Qing Wu,et al.  A low-power memristive neuromorphic circuit utilizing a global/local training mechanism , 2011, The 2011 International Joint Conference on Neural Networks.

[8]  Hai Helen Li,et al.  Spintronic Memristor Through Spin-Torque-Induced Magnetization Motion , 2009, IEEE Electron Device Letters.

[9]  L. Chua.  Memristor-The missing circuit element , 1971 .

[10]  Z. Zeng,et al.  Dynamic behaviors of a class of memristor-based Hopfield networks , 2011 .

[11]  Leon O. Chua.  Resistance switching memories are memristors , 2011 .

[12]  John J. Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities , 1999 .

[13]  Robert Proulx,et al.  NDRAM: nonlinear dynamic recurrent associative memory for learning bipolar and nonbipolar correlated patterns , 2005, IEEE Transactions on Neural Networks.

[14]  Guodong Zhang,et al.  Global exponential stability of a class of memristor-based recurrent neural networks with time-varying delays , 2012, Neurocomputing.

[15]  Zhigang Zeng,et al.  Synchronization control of a class of memristor-based recurrent neural networks , 2012, Inf. Sci..

[16]  Quan Yin,et al.  Global exponential periodicity and stability of a class of memristor-based recurrent neural networks with multiple delays , 2013, Inf. Sci..

[17]  Cong Xu,et al.  Impact of process variations on emerging memristor , 2010, Design Automation Conference.

[18]  Hui Wang,et al.  Memristor-based synapse design and training scheme for neuromorphic computing architecture , 2012, The 2012 International Joint Conference on Neural Networks (IJCNN).