Abstract
Neuromorphic Very Large Scale Integration (VLSI) devices emulate the activation dynamics of biological neuronal networks using either mixed-signal analog/digital or purely digital electronic circuits. Using analog circuits in silicon to physically emulate the functionality of biological neurons and synapses enables faithful modeling of neural and synaptic dynamics at ultra low power consumption in real-time, and thus may serve as computational substrate for a new generation of efficient neural controllers for artificial intelligent systems. Although one of the main advantages of neural networks is their ability to perform on-line learning, only a small number of neuromorphic hardware devices implement this feature on-chip. In this work, we use a reconfigurable on-line learning spiking (ROLLS) neuromorphic processor chip to build a neuronal architecture for sequence learning. The proposed neuronal architecture uses the attractor properties of winner-takes-all (WTA) dynamics to cope with mismatch and noise in the ROLLS analog computing elements, and it uses its on-chip plasticity features to store sequences of states. We demonstrate, with a proof-of-concept feasibility study how this architecture can store, replay, and update sequences of states, induced by external inputs. Controlled by the attractor dynamics and an explicit destabilizing signal, the items in a sequence can last for varying amounts of time and thus reliable sequence learning and replay can be robustly implemented in a real sensorimotor system.
Highlights
Mixed-signal analog-digital neuromorphic Very Large Scale Integration (VLSI) systems emulate the biophysics of cortical neurons and synaptic connections between them using the physics of silicon electronic devices (Moradi et al, 2018)
The architecture is presented with a sequence of items A-BC and stores them in plastic synapses that connect ordinal groups (I., II., III.) to the content Dynamic Neural Fields (DNFs)
In a proof-of-concept demonstration we have shown how sequences can be stored in a mixed signal analog/digital neuromorphic device with on-chip plasticity (Qiao et al, 2015)
Summary
Mixed-signal analog-digital neuromorphic Very Large Scale Integration (VLSI) systems emulate the biophysics of cortical neurons and synaptic connections between them using the physics of silicon electronic devices (Moradi et al, 2018). While some recent neuromorphic hardware devices aim at speeding up the processing time of computational neuroscience simulations, e.g., SpiNNaker (Furber et al, 2012) or HICANN (Schemmel et al, 2010; Benjamin et al, 2014), other neuromorphic hardware systems have been developed as basic research tools for emulating the properties of real cortical circuits in real-time (Chicca et al, 2014; Qiao et al, 2015). Neuromorphic systems that have the ability to modify synaptic weights between neurons with biologically plausible plasticity mechanism allow the construction of low-power adaptive neural processing systems that can be used for building autonomous cognitive agents (Chicca et al, 2014)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have