Hippocampus-inspired neural network runs efficiently on memristor hardware
Scaled Hippocampus-inspired Neural Networks on Neuromorphic Memristive Hardware
Neural and Evolutionary ComputingEmerging Technologies
Summary
Learning from how a part of the brain called the hippocampus works, the authors built a small but powerful neural network that mimics its structure and activity. They shrunk down the brain-inspired model to fit real hardware made up of memristors, which store information like brain synapses. Surprisingly, the physical hardware even outperformed the computer simulation by using natural noise to keep the network active. This work shows how brain-based designs paired with new hardware can lead to more energy-efficient computing.
What this means in practice
- •For neuromorphic hardware engineers: Run biologically inspired spiking neural networks efficiently on memristor-based platforms with diverse neuron types and robust dynamics.
- •For embedded systems developers: Implement energy-efficient neural models with complex dynamics on FPGA and memristor hybrid hardware for real-time computing tasks.
Authors
Joseph A. Kilgore, Jeffrey D. Kopsick, Zahin Ahmed, Giorgio A. Ascoli, Gina C. Adam
Abstract
The hippocampus, a key brain region for learning and memory, exhibits rich structural diversity, sparse communication, and robust dynamics with incredible energy efficiency. It offers promising insights for novel computing capabilities, particularly when co-designed with emerging hardware technologies. In this work, we draw inspiration from the rodent CA3 hippocampal subregion to develop the first spiking neural network with neuronal diversity and biologically-realistic resting state dynamics demonstrated on memristor hardware. We propose a network downscaling methodology utilizing a 4-prong objective function and demonstrate a small-scale CA3-inspired network with 179 Izhikevich-modeled neurons, 3 neuronal types and 17,996 synapses with similar resting-state dynamics as the orders-of-magnitude larger full-scale network. The small-scale network is mapped to an FPGA/memristor platform using a greedy algorithm and 18,316 memristors. Benefiting from memristor noise, the hardware implementation shows continuous periodic behavior, outperforming simulated hardware. This work showcases the potential of biologically-realistic algorithms on emerging hardware for neuromorphic computing.