Academic
Publications
Spike-timing-dependent learning in memristive nanodevices

Spike-timing-dependent learning in memristive nanodevices,10.1109/NANOARCH.2008.4585796,G. S. Snider

Spike-timing-dependent learning in memristive nanodevices   (Citations: 21)
BibTex | RIS | RefWorks Download
The neuromorphic paradigm is attractive for nanoscale computation because of its massive parallelism, potential scalability, and inherent defect-, fault-, and failure-tolerance. We show how to implement timing-based learning laws, such as spike-timing-dependent plasticity (STDP), in simple, memristive nanodevices, such as those constructed from certain metal oxides. Such nano-scale ldquosynapsesrdquo can be combined with CMOS ldquoneuronsrdquo to create neuromorphic hardware several orders of magnitude denser than is possible in conventional CMOS. The key ideas are: (1) to factor out two synaptic state variables to pre- and post-synaptic neurons; and (2) to separate computational communication from learning by time-division multiplexing of pulse-width-modulated signals through synapses. This approach offers the advantages of: better control over power dissipation; fewer constraints on the design of memristive materials used for nanoscale synapses; learning dynamics can be dynamically turned on or off (e.g. by attentional priming mechanisms communicated extra-synaptically); greater control over the precise form and timing of the STDP equations; the ability to implement a variety of other learning laws besides STDP; better circuit diversity since the approach allows different learning laws to be implemented in different areas of a single chip using the same memristive material for all synapses.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...Memristive synapses have been interfaced with CMOS neurons in [3] and a cellular non-linear network based on memristors is proposed in [4]...

    Jeyavijayan Rajendranet al. An Approach to Tolerate Process Related Variations in Memristor-Based ...

    • ...This crossbar architecture promises a number of advantages for neuro-inspired learning circuit design: i) It allows parallel learning [19], therefore greatly improving the learning speed, which is a strong limitation in conventional neuromorphic circuits based on serial learning [17], [20]...

    Si-Yu Liaoet al. Design and Modeling of a Neuro-Inspired Learning Circuit Using Nanotub...

    • ...In particular it has been suggested [8],[9],[15], and shown experimentally [4], that such devices could indeed reproduce a feature of biological synapses – spike timing dependent plasticity (STDP) [16],[17] – that is believed to be a foundation for learning in the brain [18],[19]...
    • ...As mentioned above, several proposals exist to use memristive devices for STDP [8],[9],[15]...
    • ...When applied a voltage pulse smaller than a negative threshold VT-, they decrease their conductance [4],[8]...
    • ...This was proposed in [9], itself close to the proposal [8] and experimentally demonstrated in [4]...

    Damien Querliozet al. Simulation of a memristor-based spiking neural network immune to devic...

    • ...STDP learning using memristive synapses, notably [23-26]...
    • ...The main difference between [23-24] and [25-26] is that in [23-24], the two-part spike is implemented as a discrete-time stepwise waveform approximation, whereas [25-26] use values calculated from continuous waveform equations, allowing them to operate in continuous time...
    • ...The main difference between [23-24] and [25-26] is that in [23-24], the two-part spike is implemented as a discrete-time stepwise waveform approximation, whereas [25-26] use values calculated from continuous waveform equations, allowing them to operate in continuous time...
    • ...We follow [23-24] in using discrete-time stepwise waveforms, as our SNNs operate in discrete time...

    Gerard Howardet al. Towards evolving spiking networks with memristive synapses

    • ...Recently, it was experimentally proved that hybrid synapse/neuron networks made up of complementary metaloxide-semiconductor (CMOS) neurons and nanoscale memristor synapses are able to support spike-timing-dependent plasticity, an important synaptic adaptation rule for competitive Hebbian learning ([5], [6], [7])...

    Fernando Corintoet al. Memristor synaptic dynamics' influence on synchronous behavior of two ...

Sort by: