Large developing receptive fields using a distributed and locally reprogrammable address-event receiver

Toward Robust Cognitive 3D Brain-Inspired Cross-Paradigm System

Spiking Neuromorphic systems have been introduced as promising platforms for energy-efficient spiking neural network (SNNs) execution. SNNs incorporate neuronal and synaptic states in addition to the variant time scale into their computational model. Since each neuron in these networks is connected to many others, high bandwidth is required. Moreover, since the spike times are used to encode information in SNN, a precise communication latency is also needed, although SNN is tolerant to the spike delay variation in some limits when it is seen as a whole. The two-dimensional packet-switched network-on-chip was proposed as a solution to provide a scalable interconnect fabric in large-scale spike-based neural networks. The 3D-ICs have also attracted a lot of attention as a potential solution to resolve the interconnect bottleneck. Combining these two emerging technologies provides a new horizon for IC design to satisfy the high requirements of low power and small footprint in emerging AI applications. Moreover, although fault-tolerance is a natural feature of biological systems, integrating many computation and memory units into neuromorphic chips confronts the reliability issue, where a defective part can affect the overall system's performance. This paper presents the design and simulation of R-NASH-a reliable three-dimensional digital neuromorphic system geared explicitly toward the 3D-ICs biological brain's three-dimensional structure, where information in the network is represented by sparse patterns of spike timing and learning is based on the local spike-timing-dependent-plasticity rule. Our platform enables high integration density and small spike delay of spiking networks and features a scalable design. R-NASH is a design based on the Through-Silicon-Via technology, facilitating spiking neural network implementation on clustered neurons based on Network-on-Chip. We provide a memory interface with the host CPU, allowing for online training and inference of spiking neural networks. Moreover, R-NASH supports fault recovery with graceful performance degradation.

Neuromorphic Spiking Neural Networks and Their Memristor-CMOS Hardware Implementations

Inspired by biology, neuromorphic systems have been trying to emulate the human brain for decades, taking advantage of its massive parallelism and sparse information coding. Recently, several large-scale hardware projects have demonstrated the outstanding capabilities of this paradigm for applications related to sensory information processing. These systems allow for the implementation of massive neural networks with millions of neurons and billions of synapses. However, the realization of learning strategies in these systems consumes an important proportion of resources in terms of area and power. The recent development of nanoscale memristors that can be integrated with Complementary Metal-Oxide-Semiconductor (CMOS) technology opens a very promising solution to emulate the behavior of biological synapses. Therefore, hybrid memristor-CMOS approaches have been proposed to implement large-scale neural networks with learning capabilities, offering a scalable and lower-cost alternative to existing CMOS systems.

Hierarchical Address Event Routing for Reconfigurable Large-Scale Neuromorphic Systems

We present a hierarchical address-event routing (HiAER) architecture for scalable communication of neural and synaptic spike events between neuromorphic processors, implemented with five Xilinx Spartan-6 field-programmable gate arrays and four custom analog neuromophic integrated circuits serving 262k neurons and 262M synapses. The architecture extends the single-bus address-event representation protocol to a hierarchy of multiple nested buses, routing events across increasing scales of spatial distance. The HiAER protocol provides individually programmable axonal delay in addition to strength for each synapse, lending itself toward biologically plausible neural network architectures, and scales across a range of hierarchies suitable for multichip and multiboard systems in reconfigurable large-scale neuromorphic systems. We show approximately linear scaling of net global synaptic event throughput with number of routing nodes in the network, at 3.6×10⁷ synaptic events per second per 16k-neuron node in the hierarchy.

A 1.5 ns OFF/ON switching-time voltage-mode LVDS driver/receiver pair for asynchronous AER bit-serial chip grid links with up to 40 times event-rate dependent power savings

This paper presents a low power fast ON/OFF switchable voltage mode implementation of a driver/receiver pair intended to be used in high speed bit-serial Low Voltage Differential Signaling (LVDS) Address Event Representation (AER) chip grids, where short (like 32-bit) sparse data packages are transmitted. Voltage-Mode drivers require intrinsically half the power of their Current-Mode counterparts and do not require Common-Mode Voltage Control. However, for fast ON/OFF switching a special high-speed voltage regulator is required which needs to be kept ON during data pauses, and hence its power consumption must be minimized, resulting in tight design constraints. A proof-of-concept chip test prototype has been designed and fabricated in low-cost standard 0.35 μ m CMOS. At ± 500 mV voltage swing with 500 Mbps serial bit rate and 32 bit events, current consumption scales from 15.9 mA (7.7 mA for the driver and 8.2 mA for the receiver) at 10 Mevent/s rate to 406 μ A ( 343 μ A for the driver and 62.5 μA for the receiver) for an event rate below 10 Kevent/s, therefore achieving a rate dependent power saving of up to 40 times, while keeping switching times at 1.5 ns. Maximum achievable event rate was 13.7 Meps at 638 Mbps serial bit rate. Additionally, differential voltage swing is tunable, thus allowing further power reductions.

Multicasting mesh AER: a scalable assembly approach for reconfigurable neuromorphic structured AER systems. Application to ConvNets

This paper presents a modular, scalable approach to assembling hierarchically structured neuromorphic Address Event Representation (AER) systems. The method consists of arranging modules in a 2D mesh, each communicating bidirectionally with all four neighbors. Address events include a module label. Each module includes an AER router which decides how to route address events. Two routing approaches have been proposed, analyzed and tested, using either destination or source module labels. Our analyses reveal that depending on traffic conditions and network topologies either one or the other approach may result in better performance. Experimental results are given after testing the approach using high-end Virtex-6 FPGAs. The approach is proposed for both single and multiple FPGAs, in which case a special bidirectional parallel-serial AER link with flow control is exploited, using the FPGA Rocket-I/O interfaces. Extensive test results are provided exploiting convolution modules of 64 × 64 pixels with kernels with sizes up to 11 × 11, which process real sensory data from a Dynamic Vision Sensor (DVS) retina. One single Virtex-6 FPGA can hold up to 64 of these convolution modules, which is equivalent to a neural network with 262 × 10(3) neurons and almost 32 million synapses.

Silicon-based dynamic synapse with depressing response

A compact implementation of a dynamic charge transfer synapse cell, capable of implementing synaptic depression, is presented. The cell is combined with a simple current mirror summing node to produce biologically plausible postsynaptic potentials (PSPs). A single charge packet is effectively transferred from the synapse to the summing node, whenever a presynaptic pulse is applied to one of its terminals. The charge packet is "weighted" by a voltage applied to the second terminal of the synapse. A voltage applied to the third terminal determines the charge recovery time in the synapse, which can be adjusted over several orders of magnitude. This voltage determines the paired pulse ratio for the synapse. The fall time of the PSP is also adjustable and is set by the gate voltage of a metal-oxide-semiconductor field-effect transistor operating in subthreshold. Results extracted from chips fabricated in a 0.35-μm complementary metal-oxide-semiconductor process, alongside theoretical and simulation results, confirm the ability of the cell to produce PSPs that are characteristic of real synapses. The concept addresses a key requirement for scalable hardware neural networks.

Adaptive visual and auditory map alignment in barn owl superior colliculus and its neuromorphic implementation

Adaptation is one of the most important phenomena in biology. A young barn owl can adapt to imposed environmental changes, such as artificial visual distortion caused by wearing a prism. This adjustment process has been modeled mathematically and the model replicates the sensory map realignment of barn owl superior colliculus (SC) through axonogenesis and synaptogenesis. This allows the biological mechanism to be transferred to an artificial computing system and thereby imbue it with a new form of adaptability to the environment. The model is demonstrated in a real-time robot environment. Results of the experiments are compared with and without prism distortion of vision, and show improved adaptability for the robot. However, the computation speed of the embedded system in the robot is slow. A digital and analog mixed signal very-large-scale integration (VLSI) circuit has been fabricated to implement adaptive sensory pathway changes derived from the SC model at higher speed. VLSI experimental results are consistent with simulation results.

Relaxed fault-tolerant hardware implementation of neural networks in the presence of multiple transient errors

Reliability should be identified as the most important challenge in future nano-scale very large scale integration (VLSI) implementation technologies for the development of complex integrated systems. Normally, fault tolerance (FT) in a conventional system is achieved by increasing its redundancy, which also implies higher implementation costs and lower performance that sometimes makes it even infeasible. In contrast to custom approaches, a new class of applications is categorized in this paper, which is inherently capable of absorbing some degrees of vulnerability and providing FT based on their natural properties. Neural networks are good indicators of imprecision-tolerant applications. We have also proposed a new class of FT techniques called relaxed fault-tolerant (RFT) techniques which are developed for VLSI implementation of imprecision-tolerant applications. The main advantage of RFT techniques with respect to traditional FT solutions is that they exploit inherent FT of different applications to reduce their implementation costs while improving their performance. To show the applicability as well as the efficiency of the RFT method, the experimental results for implementation of a face-recognition computationally intensive neural network and its corresponding RFT realization are presented in this paper. The results demonstrate promising higher performance of artificial neural network VLSI solutions for complex applications in faulty nano-scale implementation environments.

Spike-timing-dependent plasticity with weight dependence evoked from physical constraints

Analogue and mixed-signal VLSI implementations of Spike-Timing-Dependent Plasticity (STDP) are reviewed. A circuit is presented with a compact implementation of STDP suitable for parallel integration in large synaptic arrays. In contrast to previously published circuits, it uses the limitations of the silicon substrate to achieve various forms and degrees of weight dependence of STDP. It also uses reverse-biased transistors to reduce leakage from a capacitance representing weight. Chip results are presented showing: various ways in which the learning rule may be shaped; how synaptic weights may retain some indication of their learned values over periods of minutes; and how distributions of weights for synapses convergent on single neurons may shift between more or less extreme bimodality according to the strength of correlational cues in their inputs.

Robust Working Memory in an Asynchronously Spiking Neural Network Realized with Neuromorphic VLSI

We demonstrate bistable attractor dynamics in a spiking neural network implemented with neuromorphic VLSI hardware. The on-chip network consists of three interacting populations (two excitatory, one inhibitory) of leaky integrate-and-fire (LIF) neurons. One excitatory population is distinguished by strong synaptic self-excitation, which sustains meta-stable states of “high” and “low”-firing activity. Depending on the overall excitability, transitions to the “high” state may be evoked by external stimulation, or may occur spontaneously due to random activity fluctuations. In the former case, the “high” state retains a “working memory” of a stimulus until well after its release. In the latter case, “high” states remain stable for seconds, three orders of magnitude longer than the largest time-scale implemented in the circuitry. Evoked and spontaneous transitions form a continuum and may exhibit a wide range of latencies, depending on the strength of external stimulation and of recurrent synaptic excitation. In addition, we investigated “corrupted” “high” states comprising neurons of both excitatory populations. Within a “basin of attraction,” the network dynamics “corrects” such states and re-establishes the prototypical “high” state. We conclude that, with effective theoretical guidance, full-fledged attractor dynamics can be realized with comparatively small populations of neuromorphic hardware neurons.