Search

This paper presents a reconfigurable digital implementation of an event-based binaural cochlear system on a Field Programmable Gate Array (FPGA). It consists of a pair of the Cascade of Asymmetric Resonators with Fast Acting Compression (CAR FAC) cochlea models and leaky integrate and fire (LIF) neurons. Additionally, we propose an event-driven SpectroTemporal Receptive Field (STRF) Feature Extraction using Adaptive Selection Thresholds (FEAST). It is tested on the TIDIGTIS benchmark and compared with current event-based auditory signal processing approaches and neural networks., 12 pages, 8 figures

The field of neuromorphic computing holds great promise in terms of advancing computing efficiency and capabilities by following brain-inspired principles. However, the rich diversity of techniques employed in neuromorphic research has resulted in a lack of clear standards for benchmarking, hindering effective evaluation of the advantages and strengths of neuromorphic methods compared to traditional deep-learning-based methods. This paper presents a collaborative effort, bringing together members from academia and the industry, to define benchmarks for neuromorphic computing: NeuroBench. The goals of NeuroBench are to be a collaborative, fair, and representative benchmark suite developed by the community, for the community. In this paper, we discuss the challenges associated with benchmarking neuromorphic solutions, and outline the key features of NeuroBench. We believe that NeuroBench will be a significant step towards defining standards that can unify the goals of neuromorphic computing and drive its technological progress. Please visit neurobench.ai for the latest updates on the benchmark tasks and metrics.

This paper presents an efficient hardware implementation of the recently proposed Optimized Deep Event-driven Spiking Neural Network Architecture (ODESA). ODESA is the first network to have end-to-end multi-layer online local supervised training without using gradients and has the combined adaptation of weights and thresholds in an efficient hierarchical structure. This research shows that the network architecture and the online training of weights and thresholds can be implemented efficiently on a large scale in hardware. The implementation consists of a multi-layer Spiking Neural Network (SNN) and individual training modules for each layer that enable online self-learning without using back-propagation. By using simple local adaptive selection thresholds, a Winner-Takes-All (WTA) constraint on each layer, and a modified weight update rule that is more amenable to hardware, the trainer module allocates neuronal resources optimally at each layer without having to pass high-precision error measurements across layers. All elements in the system, including the training module, interact using event-based binary spikes. The hardware-optimized implementation is shown to preserve the performance of the original algorithm across multiple spatial-temporal classification problems with significantly reduced hardware requirements., Comment: 15 Pages, 19 Figures

Deep Neural Network (DNN) based inference at the edge is challenging as these compute and data-intensive algorithms need to be implemented at low cost and low power while meeting the latency constraints of the target applications. Sparsity, in both activations and weights inherent to DNNs, is a key knob to leverage. In this paper, we present RAMAN, a Re-configurable and spArse tinyML Accelerator for infereNce on edge, architected to exploit the sparsity to reduce area (storage), power as well as latency. RAMAN can be configured to support a wide range of DNN topologies - consisting of different convolution layer types and a range of layer parameters (feature-map size and the number of channels). RAMAN can also be configured to support accuracy vs power/latency tradeoffs using techniques deployed at compile-time and run-time. We present the salient features of the architecture, provide implementation results and compare the same with the state-of-the-art. RAMAN employs novel dataflow inspired by Gustavson's algorithm that has optimal input activation (IA) and output activation (OA) reuse to minimize memory access and the overall data movement cost. The dataflow allows RAMAN to locally reduce the partial sum (Psum) within a processing element array to eliminate the Psum writeback traffic. Additionally, we suggest a method to reduce peak activation memory by overlapping IA and OA on the same memory space, which can reduce storage requirements by up to 50%. RAMAN was implemented on a low-power and resource-constrained Efinix Ti60 FPGA with 37.2K LUTs and 8.6K register utilization. RAMAN processes all layers of the MobileNetV1 model at 98.47 GOp/s/W and the DS-CNN model at 79.68 GOp/s/W by leveraging both weight and activation sparsity.

As an emerging approach to space situational awareness and space imaging, the practical use of an event-based camera in space imaging for precise source analysis is still in its infancy. The nature of event-based space imaging and data collection needs to be further explored to develop more effective event-based space image systems and advance the capabilities of event-based tracking systems with improved target measurement models. Moreover, for event measurements to be meaningful, a framework must be investigated for event-based camera calibration to project events from pixel array coordinates in the image plane to coordinates in a target resident space object's reference frame. In this paper, the traditional techniques of conventional astronomy are reconsidered to properly utilise the event-based camera for space imaging and space situational awareness. This paper presents the techniques and systems used for calibrating an event-based camera for reliable and accurate measurement acquisition. These techniques are vital in building event-based space imaging systems capable of real-world space situational awareness tasks. By calibrating sources detected using the event-based camera, the spatio-temporal characteristics of detected sources or `event sources' can be related to the photometric characteristics of the underlying astrophysical objects. Finally, these characteristics are analysed to establish a foundation for principled processing and observing techniques which appropriately exploit the capabilities of the event-based camera.

Many models of evolution are stochastic processes, where some quantity of interest fluctuates randomly in time. One classic example is the Moran birth-death process, where that quantity is the number of mutants in a population. In such processes we are often interested in their absorption (i.e. fixation) probabilities, and the conditional distributions of absorption time. Those conditional time distributions can be very difficult to calculate, even for relatively simple processes like the Moran birth-death model. Instead of considering the time to absorption, we consider a closely-related quantity: the number of mutant population size changes before absorption. We use Wald’s martingale to obtain the conditional characteristic functions of that quantity in the Moran process. Our expressions are novel, analytical, and exact. The parameter dependence of the characteristic functions is explicit, so it is easy to explore their properties in parameter space. We also use them to approximate the conditional characteristic functions of absorption time. We state the conditions under which that approximation is particularly accurate. Martingales are an elegant framework to solve principal problems of evolutionary stochastic processes. They do not require us to evaluate recursion relations, so we can quickly and tractably obtain absorption probabilities and times of evolutionary stochastic processes.Author summaryThe Moran process is a probabilistic birth-death model of evolution. A mutant is introduced to an indigenous population, and we randomly choose organisms to live or die on subsequent time steps. Our goals are to calculate the probabilities that the mutant eventually dominates the population or goes extinct, and the distribution of time it requires to do so. The conditional distributions of time are difficult to obtain for the Moran process, so we consider a slightly different but related problem. We instead calculate the conditional distributions of the number of times that the mutant population size changes before it dominates the population or goes extinct. We use a martingale identified by Abraham Wald to obtain elegant and exact expressions for those distributions. We then use them to approximate conditional time distributions, and we show when that approximation is accurate. Our analysis outlines the basic concepts martingales and demonstrates why they are a formidable tool for studying probabilistic evolutionary models such as the Moran process.

Unsupervised feature extraction algorithms form one of the most important building blocks in machine learning systems. These algorithms are often adapted to the event-based domain to perform online learning in neuromorphic hardware. However, not designed for the purpose, such algorithms typically require significant simplification during implementation to meet hardware constraints, creating trade offs with performance. Furthermore, conventional feature extraction algorithms are not designed to generate useful intermediary signals which are valuable only in the context of neuromorphic hardware limitations. In this work a novel event-based feature extraction method is proposed that focuses on these issues. The algorithm operates via simple adaptive selection thresholds which allow a simpler implementation of network homeostasis than previous works by trading off a small amount of information loss in the form of missed events that fall outside the selection thresholds. The behavior of the selection thresholds and the output of the network as a whole are shown to provide uniquely useful signals indicating network weight convergence without the need to access network weights. A novel heuristic method for network size selection is proposed which makes use of noise events and their feature representations. The use of selection thresholds is shown to produce network activation patterns that predict classification accuracy allowing rapid evaluation and optimization of system parameters without the need to run back-end classifiers. The feature extraction method is tested on both the N-MNIST benchmarking dataset and a dataset of airplanes passing through the field of view. Multiple configurations with different classifiers are tested with the results quantifying the resultant performance gains at each processing stage., Comment: 15 Pages. 9 Figures

BACKGROUND: Peripheral neuropathic desensitization associated with aging, diabetes, alcoholism and HIV/AIDS, affects tens of millions of people worldwide, and there is little or no treatment available to improve sensory function. Recent studies that apply imperceptible continuous vibration or electrical stimulation have shown promise in improving sensitivity in both diseased and healthy participants. This class of interventions only has an effect during application, necessitating the design of a wearable device for everyday use. We present a circuit that allows for a low-power, low-cost and small form factor implementation of a current stimulator for the continuous application of subthreshold currents. RESULTS: This circuit acts as a voltage-to-current converter and has been tested to drive + 1 to - 1 mA into a 60 k[Formula: see text] load from DC to 1 kHz. Driving a 60 k[Formula: see text] load with a 2 mA peak-to-peak 1 kHz sinusoid, the circuit draws less than 21 mA from a 9 V source. The minimum operating current of the circuit is less than 12 mA. Voltage compliance is ± 60 V with just 1.02 mA drawn by the high voltage current drive circuitry. The circuit was implemented as a compact 46 mm × 21 mm two-layer PCB highlighting its potential for use in a body-worn device. CONCLUSIONS: No design to the best of our knowledge presents comparably low quiescent power with such high voltage compliance. This makes the design uniquely appropriate for low-power transcutaneous current stimulation in wearable applications. Further development of driving and instrumentation circuitry is recommended.

Whilst most engineered systems use signals that are continuous in time, there is a domain of systems in which signals consist of events. Events, like Dirac delta functions, have no meaningful time duration. Many important real-world systems are intrinsically event-based, including the mammalian brain, in which the primary packets of data are spike events, or action potentials. In this domain, signal processing requires responses to spatio-temporal patterns of events. We show that some straightforward modifications to the standard ELM topology produce networks that are able to perform spatio-temporal event processing online with a high degree of accuracy. The modifications involve the re-definition of hidden layer units as synaptic kernels, in which the input delta functions are transformed into continuous-valued signals using a variety of impulse-response functions. This permits the use of linear solution methods in the output layer, which can produce events as output, if modeled as a classifier; the output classes are 'event' or 'no event'. We illustrate the method in application to a spike-processing problem., Comment: Accepted for publication in Neurocomputing

We present an alternative to the pseudo-inverse method for determining the hidden to output weight values for Extreme Learning Machines performing classification tasks. The method is based on linear discriminant analysis and provides Bayes optimal single point estimates for the weight values., Comment: In submission to the ELM 2014 conference

INI's posters