Gast ::
| Hauptseite > Publikationsdatenbank > Sequence learning in a spiking neural network with memristive synapses |
| Hauptseite > Publikationsdatenbank > Sequence learning in a spiking neural network with memristive synapses |
| Poster (After Call) | FZJ-2022-01972 |
; ; ; ; ;
2022
Abstract: Brain-inspired computing proposes a set of algorithmic principles that hold promise for advancing artificial intelligence. They endow systems with self-learning capabilities, efficient energy usage, and high storage capacity. A core concept that lies at the heart of brain computation is sequence learning and prediction. This form of computation is essential for almost all our daily tasks such as movement generation, perception, and language. Understanding how the brain performs such a computation is not only important to advance neuroscience but also to pave the way to new technological brain-inspired applications. A previously developed spiking neural network implementation of sequence prediction and recall learns complex, high-order sequences in an unsupervised manner by local, biologically inspired plasticity rules. An emerging type of hardware that holds promise for efficiently running this type of algorithm is analog neuromorphic hardware (ANH). It emulates the brain architecture and maps neurons and synapses directly into a physical substrate. Memristive devices have been identified as potential synaptic elements in ANH. In particular, redox-induced resistive random access memories (ReRAM) devices stand out at many aspects. They permit scalability, are energy-efficient and fast, and can implement biological learning rules. In this work, we study the feasibility of using ReRAM devices as a replacement of the biological synapses in the sequence learning model. We implement and simulate the model including the ReRAM plasticity using the neural simulator NEST. We investigate two types of ReRAM devices: (i) an analog switching memristive device, where the conductance gradually changes between a low conductance (LCS) and a high conductance state (HCS), and (ii) a binary switching memristive device, where the conductance abruptly changes between the LCS and the HCS. We study the performance characteristics of the sequence learning model as a function of different device properties, and demonstrate resilience with respect to different on/off ratios, conductance resolutions, device variability, and synaptic failure.
|
The record appears in these collections: |