Structural plasticity on an accelerated analog neuromorphic hardware system
This work addresses resource constraints in neuromorphic computing for computational neuroscience and machine learning, representing an incremental improvement in hardware optimization.
The authors tackled the problem of limited neural connectivity and synaptic capacity in neuromorphic hardware by developing a structural plasticity algorithm that rewires connections while maintaining constant fan-in and sparsity, implemented on the BrainScaleS-2 system and shown to optimize network topology and computational efficiency in a supervised learning scenario.
In computational neuroscience, as well as in machine learning, neuromorphic devices promise an accelerated and scalable alternative to neural network simulations. Their neural connectivity and synaptic capacity depends on their specific design choices, but is always intrinsically limited. Here, we present a strategy to achieve structural plasticity that optimizes resource allocation under these constraints by constantly rewiring the pre- and gpostsynaptic partners while keeping the neuronal fan-in constant and the connectome sparse. In particular, we implemented this algorithm on the analog neuromorphic system BrainScaleS-2. It was executed on a custom embedded digital processor located on chip, accompanying the mixed-signal substrate of spiking neurons and synapse circuits. We evaluated our implementation in a simple supervised learning scenario, showing its ability to optimize the network topology with respect to the nature of its training data, as well as its overall computational efficiency.