Structural plasticity for neuromorphic networks with electropolymerized dendritic PEDOT connections

Abstract

Neural networks are powerful tools for solving complex problems, but finding the right network topology for a given task remains an open question. Biology uses neurogenesis and structural plasticity to solve this problem. Advanced neural network algorithms are mostly relying on synaptic plasticity and learning. The main limitation in reconciling these two approaches is the lack of a viable hardware solution that could reproduce the bottom-up development of biological neural networks. Here, we show how the dendritic growth of PEDOT:PSS-based fibers through AC electropolymerization can implement structural plasticity during network development. We find that this strategy follows Hebbian principles and is able to define topologies that leverage better computing performances with sparse synaptic connectivity for solving non-trivial tasks. This approach is validated in software simulation, and offers up to 61% better network sparsity on classification and 50% in signal reconstruction tasks.

Graphene-Based Artificial Dendrites for Bio-Inspired Learning in Spiking Neuromorphic Systems

“…Supervised learning was chosen as the application because it is the nearest-term application for analog accelerators due to high accuracy compared to that of training using unsupervised methods. It is important to note that dendritic behavior has been shown to have benefits when applied to bio-inspired learning algorithms shown in refs , , , and and it is an important building block for biological systems. The Fashion-MNIST clothing article classification task was chosen as the benchmark.…”

“…Previous work on analog neuromorphic devices for artificial dendrites has focused on using CMOS, requiring circuit overhead when designing a neuronal configuration with multiple dendrites. , Other works have proposed the use of emerging memories as artificial dendrites. For example, PEDOT has been shown to electropolymerize and form synaptic connections between dendritic branches that can be controlled through unsupervised learning methods, though the potentiation of the dendritic branches was not characterized . Silicon nanowire, ionic nanomemristor, and starch-based transistor artificial synapses have been shown to benefit from the dendrite-like physical dynamics of the ion-gated devices.…”