Spiking neural networks (SNNs) are well suited for spatio-temporal learning and implementations on energy-efficient event-driven neuromorphic processors. However, existing SNN error backpropagation (BP) methods lack proper handling of spiking discontinuities suffer from low performance compared with the BP traditional artificial networks. In addition, a large number time steps typically required to achieve decent performance, leading high latency rendering spike-based computation unscalable...

Spiking neural networks (SNNs) well support spatiotemporal learning and energy-efficient event-driven hardware neuromorphic processors. As an important class of SNNs, recurrent spiking (RSNNs) possess great computational power. However, the practical application RSNNs is severely limited by challenges in training. Biologically-inspired unsupervised has capability boosting performance RSNNs. On other hand, existing backpropagation (BP) methods suffer from high complexity unrolling time,...

Rail transportation is used extensively in urban areas to reduce CO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> emission and increase road capacity. As a result, the energy efficiency of rail becoming popular research topic. Energy-efficient train operation involves four types control: maximal traction, cruising, coasting, braking. With rapid development storage devices (ESDs), this paper aims develop an integrated optimization model...

The accumulation of aniline in the natural environment poses a potential threat to crops, and thus, investigating effects on plants holds practical implications for agricultural engineering its affiliated industries. This study combined physiological, transcriptomic, metabolomic methods investigate growth status molecular-level response mechanisms rice under stress from varying concentrations aniline. At concentration 1 mg/L, exhibited slight growth-promoting effect rice. However, higher...

Spiking neural networks (SNNs) present a promising computing model and enable bio-plausible information processing event-driven based ultra-low power neuromorphic hardware. However, training SNNs to reach the same performances of conventional deep artificial (ANNs), particularly with error backpropagation (BP) algorithms, poses significant challenge due inherent complex dynamics non-differentiable spike activities spiking neurons. In this paper, we first study on realizing competitive...

Aniline is widely used in the fields of industry and agriculture. pollution emerging as a serious global environmental problem, whereas toxicity stress mechanism on higher plants unclear. In order to clarify growth, oxidative DNA damage caused by aniline plants, we cultured two varieties rice different concentration solution. The results demonstrated that germination rate, amylase activity lipase were significantly inhibited during period when was ≥25 mg L−1 (p < 0.05 or p 0.01). Besides,...

Spiking Neural Networks (SNNs) are brain- inspired computing models incorporating unique temporal dynamics and event-driven processing. Rich in both space time offer great challenges opportunities for efficient processing of sparse spatiotemporal data compared with conventional artificial neural networks (ANNs). Specifically, the additional overheads handling added dimension limit computational capabilities neuromorphic accelerators. Iterative at every time-point inputs a temporally...

Spiking neural networks (SNNs) are the third generation of and can explore both rate temporal coding for energy-efficient event-driven computation. However, decision accuracy existing SNN designs is contingent upon processing a large number spikes over long period. Nevertheless, switching power hardware accelerators proportional to processed while length spike trains limits throughput static efficiency. This paper presents first study on developing compression significantly boost reduce...

Abstract Regional warming and atmospheric ozone (O 3 ) pollution are two of the most important environmental issues, commonly coexist in many areas. Both factors have an intense impact on plants. However, little information is available combined interactive effects air elevated O concentrations physiological characteristics To explore this issue, we studied variations growth, photosynthesis leaves Acer ginnala seedlings exposed to control (ambient temperature ), increasing + 2 °C),...

The current study employs a novel nonlinear robust control approach for path-following of underactuated autonomous underwater vehicles (AUVs) with multiple uncertainties in the vertical plane. Firstly, AUV model is established to characterize dynamics and error. To resolve dependence on detailed that appeared previous studies, unknown time-varying attack angular velocity dynamic error considered as kinematic uncertainty, while linear superposition external environmental disturbances,...

As an important class of spiking neural networks (SNNs), recurrent (RSNNs) possess great computational power and have been widely used for processing sequential data like audio text. However, most RSNNs suffer from two problems. First, due to the lack architectural guidance, random connectivity is often adopted, which does not guarantee good performance. Second, training in general challenging, bottlenecking achievable model accuracy. To address these problems, we propose a new type RSNN,...

In neural circuits, recurrent connectivity plays a crucial role in network function and stability. However, existing spiking networks (RSNNs) are often constructed by random connections without optimization. While RSNNs can produce rich dynamics that critical for memory formation learning, systemic architectural optimization of is still an open challenge. We aim to enable systematic design large via new scalable RSNN architecture automated compose based on layer called Sparsely-Connected...

Gametocidal (Gc) chromosomes can kill gametes that lack them by causing chromosomal breakage to ensure their preferential transmission, and they have been exploited in genetic breeding. The present study investigated the possible roles of small RNAs (sRNAs) Gc action. By sequencing two RNA libraries from anthers Triticum aestivum cv. Chinese Spring (CS) Spring-Gc 3C chromosome monosomic addition line (CS-3C), we identified 239 conserved 72 putative novel miRNAs, including 135 differentially...

Intrinsic plasticity (IP) is a non-Hebbian learning mechanism that self-adapts intrinsic parameters of each neuron as opposed to synaptic weights, offering complimentary opportunities for performance improvement. However, integrating IP onchip enable per-neuron self-adaptation can lead very large design overheads. This paper the first work exploring efficient on-chip neural accelerators based on recurrent spiking network model liquid state machine (LSM). The proposed LSM processor integrated...

With the development of deep learning-based methods, automated classification electrocardiograms (ECGs) has recently gained much attention. Although effectiveness neural networks been encouraging, lack information given by outputs restricts clinicians' reexamination. If uncertainty estimation comes along with results, cardiologists can pay more attention to "uncertain" cases. Our study aims classify ECGs rejection based on data and model uncertainty. We perform experiments a real-world...

Ventricular arrhythmias (VA) are the main causes of sudden cardiac death. Developing machine learning methods for detecting VA based on electrocardiograms (ECGs) can help save people's lives. However, developing such models ECGs is challenging because following: 1) group-level diversity from different subjects and 2) individual-level moments a single subject. In this study, we aim to solve these problems in pre-training fine-tuning stages. For stage, propose novel model agnostic...

Spiking neural networks (SNNs) are brain-inspired event-driven models of computation with promising ultra-low energy dissipation. Rich network dynamics emergent in recurrent spiking (R-SNNs) can form temporally based memory, offering great potential processing complex spatiotemporal data. However, recurrence connectivity produces tightly coupled data dependency both space and time, rendering hardware acceleration R-SNNs challenging. We present the first work to exploit parallelisms...