Energy efficient artificial gustatory system for in-sensor computing

Neuromorphic computing [1], inspired by the energy-efficient and parallelized architecture of the human brain, represents a paradigm shift in computing. This approach not only drives innovative hardware designs but also inspires novel network architectures like spiking neural networks (SNNs), which process information as discrete spike events. However, scalability in neuromorphic systems is constrained by the large hardware footprint. Realizing multimodal functionality using the same set of hardware primitives can offer a promising solution to these limitations.In this work, we present a dual encoder-neuron circuit based on a nanoporous graphene (NPG) memristor device [2]. The graphene leaky integrate-and-fire (GLIF) circuit has a versatile multimodal design that combines the roles of a LIF neuron and a spike encoder within the same architecture. This design mimics the dual functionality of biological neurons, which encode and process signals locally. The spike encoding capability of the circuit is further revealed using single and double layer SNN networks for pattern recognition tasks using the Modified National Institute of Standards and Technology (MNIST) dataset.Inset of Figure 1(a) shows the schematic of the fabricated NPG device with a channel length of 100 µm. The device has a lateral structure with NPG channel and gold (Au) electrodes. Figure 1(a) shows the current-voltage (I-V) characteristics of the fabricated NPG device which shows a threshold switching behavior with a wide hysteresis window and a threshold voltage (Vth) of 4.9 V. A behavioral SPICE model is developed based on a voltage-controlled switch model. It is seen from Figure 1(a) that the I-V curves simulated using the SPICE model matches the experimentally observed device switching characteristics. This reveals that the simplified SPICE model can be a good substitute for complex Verilog based models. Based on the SPICE model, we designed a simple neuron circuit with a resistor (Rs, 12 kΩ) in series with a parallel combination of a capacitor (Cm, 10 nF) and the NPG device (SPICE model). Figure 1(b) shows the variation of the membrane potential (Vm) and the output spike current in response to the input pulse voltage (Vin, 6 V). Vm shows a clear leaky integration behavior and full reset to zero voltage after the spike response. These are essential attributes of a bioplausible LIF neuron circuit. In order to develop the multimodal GLIF circuit, the output spike frequency of the LIF neuron circuit is plotted as a function of the Rs values. This mapping allows pixel values from MNIST input images to be directly translated into normalized conductance values (G/G0). Hence, each pixel is represented as a normalized conductance value, establishing a clear correlation between pixel intensity and spike frequency. This core idea is further used to develop the GLIF encoder based on the same neuron circuit as before. Based on the dual encoder-neuron circuit, we designed a single layer SNN for recognising the handwritten digits from the MNIST dataset. Figure 1(c) shows the spike-encoded images for the MNIST digits “2” and “5”. The probability of spiking in each timestep is proportional to the spiking frequency. The spike encoded images at each time step reveal the contours of the original input images using sparse representation. These sparse matrices highlight the energy efficiency and computational advantages of the GLIF system. The performance of the dual encoder-neuron system is evaluated by training the single layer SNN using both the GLIF encoder and neuron circuit as integral components. The single layer SNN recorded a high pattern recognition accuracy of 90.77% which is comparable to the purely software-based implementation (92.69%). To explore the scalability of the GLIF system, we implemented a double-layer SNN using the GLIF system, achieving a high recognition accuracy of 97.37%. The GLIF circuit exemplifies how reconfigurable primitive components can drive the development of scalable neuromorphic hardware.

Spiking Irregularity and Frequency Modulate the Behavioral Report of Single-Neuron Stimulation

Self-aware artificial auditory neuron with a triboelectric sensor for spike-based neuromorphic hardware

Spiking Neural Networks (SNN) coupled with Neuromorphic Computing is considered as a bio-plausible, power and data-efficient AI/ML technique especially suitable for resource-constrained edge devices such as robots, satellites, automated cars etc. While the performance of an SNN often suffers due to lossy spike encoding of real valued data, its implementation on actual hardware is also limited due to commercial unavailability of neuromorphic chips. While on one hand, better encoding techniques are being improvised, on the other hand, FPGAs (instead of ASICs) are nowadays considered as a plausible alternative for hardware implementation. In this paper, we have shown how a Mutual Information based information maximization technique for spike encoding can be successfully implemented on a Xilinx Zynq ZedBoard. The implemented system is found to perform at par with software simulation w a minimal error of 0.014 when tested on four different univariate time series. When running at a frequency of 200MHz, the system is observed to consume 0.018 watts (dynamic power), have a latency of 10.35 ms and a throughput of 96 samples/sec. Moreover, the FPGA-based system has the advantage of being used as a plug-and-play encoder device for any neuromorphic computing platform.

Efficient fusion of spiking neural networks and FET-type gas sensors for a fast and reliable artificial olfactory system

<span lang="EN-US">In this study, a novel neuro-fuzzy clustering algorithm is proposed based on spiking neural network and ink drop spread (IDS) concepts. The proposed structure is a one-layer artificial neural network with leaky integrate and fire (LIF) neurons. The structure implements the IDS algorithm as a fuzzy concept. Each training data will result in firing the corresponding input neuron and its neighboring neurons. A synchronous time coding algorithm is used to manage input and output neurons firing time. For an input data, one or several output neurons of the network will fire; confidence degree of the network to outputs is defined as the relative delay of the firing times with respect to the synchronous pulse. A memristor crossbar-based hardware is utilized for hardware implementation of the proposed algorithm. The simulation result corroborates that the proposed algorithm can be used as a neuro-fuzzy clustering and vector quantization algorithm.</span>

Engineering neural networks to perform specific tasks often represents a monumental challenge in determining network architecture and parameter values. In this work, we extend our previously-developed method for tuning networks of non-spiking neurons, the “Functional subnetwork approach” (FSA), to the tuning of networks composed of spiking neurons. This extension enables the direct assembly and tuning of networks of spiking neurons and synapses based on the network's intended function, without the use of global optimization or machine learning. To extend the FSA, we show that the dynamics of a generalized linear integrate and fire (GLIF) neuron model have fundamental similarities to those of a non-spiking leaky integrator neuron model. We derive analytical expressions that show functional parallels between: (1) A spiking neuron's steady-state spiking frequency and a non-spiking neuron's steady-state voltage in response to an applied current; (2) a spiking neuron's transient spiking frequency and a non-spiking neuron's transient voltage in response to an applied current; and (3) a spiking synapse's average conductance during steady spiking and a non-spiking synapse's conductance. The models become more similar as additional spiking neurons are added to each population “node” in the network. We apply the FSA to model a neuromuscular reflex pathway two different ways: Via non-spiking components and then via spiking components. These results provide a concrete example of how a single non-spiking neuron may model the average spiking frequency of a population of spiking neurons. The resulting model also demonstrates that by using the FSA, models can be constructed that incorporate both spiking and non-spiking units. This work facilitates the construction of large networks of spiking neurons and synapses that perform specific functions, for example, those implemented with neuromorphic computing hardware, by providing an analytical method for directly tuning their parameters without time-consuming optimization or learning.

This paper presents an electronic skin (e-skin) taxel array readout chip in 0.18m CMOS technology, achieving the highest reported spatial resolution of 200m, comparable to human fingertips. A key innovation is the integration on chip of a 1216 polyvinylidene fluoride (PVDF)-based piezoelectric sensor array with per-taxel signal conditioning frontend and spiking readout combined with local embedded neuromorphic first-order processing through Complex Receptive Fields (CRFs). Experimental results show that Spiking Neural Network (SNN)-based classification of the chip's spatiotemporal spiking output for input tactile stimuli such as texture and flutter frequency achieves excellent accuracies up to 97.1 and 99.2, respectively. SNN-based classification of the indentation period applied to the on-chip PVDF sensors achieved 95.5 classification accuracy, despite using only a small 256-neuron SNN classifier, a low equivalent spike encoding resolution of 3-5 bits, and a sub-Nyquist 2.2kevent/s population spiking rate, a state-of-the-art power consumption of 12.33nW per-taxel, and 75W-5mW for the entire chip is obtained. Finally, a comparison of the texture classification accuracies between two on-chip spike encoder outputs shows that the proposed neuromorphic level-crossing sampling (N-LCS) architecture with a decaying threshold outperforms the conventional bipolar level-crossing sampling (LCS) architecture with fixed threshold.

Neuromorphic computing aims at both accelerating machine learning computation and reducing its power consumption by adopting the hardware architecture of Spiking Neural Network (SNN), which is more biologically plausible than Artificial Neural Network (ANN). Rate coded SNNs, which represent signals according to the frequency of spikes, are highly compatible with ANNs and can utilize the technology accumulated in recent years. However, rate coded SNNs using a large number of spikes suffer from long latency and increased power consumption, and the search for a more efficient coding method has become active. In this study, we propose a novel coding scheme that uses not only the frequency of spikes but also the voltage amplitude and time width of spikes, which are domains that have not been used in conventional SNNs. The proposed multi-domain neural coding can be expected to increase the amount of information that can be transmitted per spike compared to ordinary rate coding. In addition, conductance-based synaptic devices/circuits like memristor have been widely studied, but current-based synaptic operations consume a lot of power. In this work, we employed memcapacitor and charge-pump based synapse circuit, which is a voltage-domain synaptic operation that ideally consumes power only during switching. We have performed a preliminary evaluation for the proposed coding scheme on a synaptic/neuron circuit designed in a ROHM 180nm CMOS process by simulating the operation on HSPICE. Moreover it was demonstrated that the proposed circuit can successfully perform synaptic operations to allow the new coding scheme.

<sec>In current sensing-computing systems, sensors are used to acquire information from environments, such data are normally analogue, unstructured and even redundant. After the analogue-to-digital conversion (ADC), the data are transferred into digital computers for processing. In computers with the von Neumann architecture, memories and central processing units (CPUs) are physically separated. Such a separation of sensing terminals, memories and CPUs yields serious problems, such as high energy consumption, long response time, huge data storage, and stringent requirements for the communication bandwidth and security. However, time- and energy-efficient ways are urgently required to process information at where data are generated. On the other hand, biological sensory organs respond to external stimuli in real-time with high efficiency due to the integrated capabilities of sensing, memory and computing. Therefore, the problem of separated sensing units, memories and processing units can be solved by emulating biological sensory organs.</sec><sec>In this work, we propose bio-inspired sensory systems with integrated capabilities of sensing, data storage and processing. In such a system, different sensors are used to capture the environmental signals from e.g. gases, light, audio and pressure, then the sensory signals are processed by an analogue signal processor, so that the energy-consuming ADC is avoided, afterwards the sensory signals are processed by a brain-inspired chip which consists of neuron-synapse cores based on memristors. In the neuron-synapse cores, leaky integrate-and-fire (LIF) neurons can be implemented by memristors and capacitors, and adaptive LIF neurons are developed from the LIF neurons to realize unsupervised learning algorithms. The synapses are realized by memristor arrays which can also perform the in-memory computing. By changing the connection between the neurons, the brain-inspired chip can realize different spiking neural networks (SNNs), such as fully connected SNN, convolutional SNN, and recurrent SNN. The synaptic weight in SNNs can be updated according to the spike-timing dependent plasticity (STDP) or the spike-rate dependent plasticity (SRDP). As an example, a bio-inspired olfactory system is demonstrated. In a artificial olfactory system, a sensor array detects and transforms the chemical information about gas molecules into electrical sensory signals. Then the sensory signals are processed by the analogue signal processing unit. After pre-processing, the brain-inspired chip classifies gases by constructing a fully connected SNN with two layers. Such a bio-inspired olfactory system emulates the function of a biological nose, overcoming the low efficiency caused by the frequent sampling, data conversion, transfer and storage under the current sensing-computing architecture. More importantly, the approach of this work can be used to emulate almost all the biological perceptions, such as touch, sight, hearing and taste, through the integration with different types of sensors., Therefore, this work offers a brand new approach to realizing the artificial intelligence (AI).</sec>

Spike timing dependent plasticity (STDP) learning rule is one of hot topics in neurobiology since it's been widely believed that synaptic plasticity mainly contribute to learning and memory in brain. Up to now, STDP has been observed in a wide variety of areas of brain, hippocampus, cortex and so on. Competition among synapses is an important behavior for this learning rule. In present study, we propose a single layer spiking neural network model using STDP learning rule in inhibitory synapses to investigate the competitive behavior. The experiments show that the synapses among neurons are both strengthened on the whole training process. Thus neurons inhibit the activities of one another, eventually the neuron with the highest input spike rate win the competition. We have found that the behavior is efficient when the differences of firing rates of input neurons without STDP are great than 5Hz, otherwise the winner neuron is random. In order to use the principle to artificial intelligent system, we use a mechanism of dynamic learning rate to let the neuron with the highest input to be selected by the competitive behavior as the winner. Therefore, a robust competitive spiking neural network is obtained.

Neuronal oscillatory phase is suggested to be associated with feature coding, carrying information for stimulus identity and neuronal activation, while phase synchronization is indicated to be correlated with signal routing, establishing flexible communication structures for neuronal interactions. Recent electrophysiological and computational studies have revealed that oscillatory phase has close relationships with neuronal excitation and input stimulus. To simulate and further investigate these issues, we simulated orientation columns with a spiking neural network and performed spectral computations according to physiological experiments. Besides, six network activity states, pre-stimulus, and stimulus periods were introduced in our simulation for both independent and comparative analyses. The simulation results demonstrated that gamma band neuronal oscillations existed in the network and even emerged during pre-stimulus period. An input stimulus orientation, if approximately preferred, could produce smaller and more concentrated oscillatory phases, but relatively stronger phase synchronization. In particular, the oscillatory phase and phase synchronization had quantifiable relationships with neuronal excitation and input strength. With the network activity state transforming gradually from strong oscillation to non-oscillation, the oscillatory phase became more and more scattered and the strength of phase synchronization declined significantly. Their relationships with neuronal excitation and input strength became increasingly unstable, and finally collapsed.

Neuromorphic computing provides a promising alternative to conventional von Neumann architectures by addressing memory bottlenecks and energy inefficiencies. In this work, a hybrid memristor‐based neuromorphic system is presented, comprising a volatile Pt/V/AlO x /Pt threshold switching memristor (TSM) as the spiking neuron and a non‐volatile Ta/ZrO x /Pt resistive switching memristor (RSM) as the artificial synapse. The neuron circuit realizes a leaky‐integrate‐and‐fire (LIF) model with tunable synaptic weights and temporal encoding functions. The artificial synapses exhibit multi‐level resistance states and retention characteristics, ensuring long‐term conductance retention. By integrating the neuron circuit with multiple artificial synapses, it is demonstrated that membrane potential (V M ) dynamics are jointly governed by input strength and timing, thereby enabling precise modulation of spike generation. In addition, the TSM‐based neuron can also serve as a spike encoder, where the output firing rate or time‐to‐first‐spike (TTFS) is directly modulated by the amplitude of the input signal. This dual functionality allows the same circuit to function as both a computational unit and an analog‐to‐spike encoder. Overall, the circuit replicates essential neuronal dynamics, including subthreshold integration, temporal summation, and synchronized firing, validating its potential for scalable neuromorphic hardware implementation.

Spiking Self-organizing Maps for Classification Problem

The present research letter is dedicated to a detailed analysis of a double-gate tunnel field-effect transistor (DG-TFET). The DG-TFET provides improved on-current (I_ON) than a conventional TFET via band-to-band (B2B) tunneling. However, DG-TFET is disadvantageous for low-power applications because of increased off-current (I_OFF) due to the large ambipolar current (I_amb). In this research work, a Si/GaAs/GaAs heterostructure DG-TFET is considered as research base for investigation of device performance.