Aug 2026· International Conference on Systems· pp. 83-90· 0 citations· 19 references
Computer Science
TL;DR
The results demonstrate that combining selective sequence modeling with spiking computation yields a sparse and efficient mechanism for state space control, establishing a principled bridge between modern SSMs and neuromorphic computing.
Abstract
While selective state space models (SSMs) have demonstrated strong performance in long-range sequence modeling, they rely on dense, continuous gating networks, limiting their deployment on event-driven neuromorphic hardware. Here, we introduce selective spiking state space models (4SM), a neuromorphic sequence modeling architecture in which selective SSM dynamics are implemented via event-driven spiking control. We show that selectivity can be realized through conditional state updates driven by sparse spiking activity. Inspired by biological synaptic integration, we propose the SelectiveSpikingGate module, which combines leaky integrate-and-fire dynamics with synaptic exponential moving average filtering, effectively bridging discrete spike events and continuous SSM dynamics. In addition, we introduce a low-rank parameterization of the selective readout, reducing parameter count by 60% while preserving expressiveness. We evaluate 4SM on permuted and sequential MNIST (psMNIST, sMNIST) as well as selective copying benchmarks, achieving competitive performance relative to non-spiking selective SSMs and outperforming prior spiking SSMs lacking selectivity. Furthermore, on the Long Range Arena (LRA) ListOps task, 4SM surpasses the gold-standard S4 baseline and significantly outperforms standard Transformer models. Our results demonstrate that combining selective sequence modeling with spiking computation yields a sparse and efficient mechanism for state space control, establishing a principled bridge between modern SSMs and neuromorphic computing.
This work proposes a noisy group neuron (NGN) model, which incorporates population-level synchronous resetting and neural stochasticity as fundamental computational mechanisms, and develops the NGN method as a framework that combines the NGN model with backpropagation learning based on mean-field dynamics.
Yajie Zhai, Yanmei Kang, Meng Li et al.· 0 citations
Spiking neural networks (SNNs) have garnered significant attention in reinforcement learning tasks for their low power consumption. However, traditional spiking reinforcement learning (SRL) methods, which rely on local-connected encoding and fixed-threshold learning, struggle to capture the inter-dimensional correlations of input information within short timesteps, limiting the network’s expressive capacity at low timesteps. While increasing timesteps can significantly enhance performance, excessive timesteps result in substantial delays. To address this contradiction and enhance the expressive and decision-making capabilities of SNNs within short timesteps, we propose Mask-Adaptive Global Connection (MAGC), a novel encoding method that efficiently captures long-range dependencies via sparse, adaptively masked connections—enabling global feature interaction in a single timestep. Additionally, dynamic-threshold spiking neurons are introduced to effectively capture and distinguish subtle changes in input signals at each timestep, thereby enhancing the spatial-temporal state representation during spike information transmission. Extensive experimental results demonstrate that the proposed method achieves performance comparable to state-of-the-art algorithms using only a single timestep, while significantly reducing inference latency and energy consumption. When extended to multiple timesteps, our approach consistently outperforms existing methods, showing substantial improvements across eight continuous control tasks from OpenAI Gym.
Rong Xiao, Zhiyuan Hu, Ping He et al.· IEEE Transactions on Image P...· 0 citations
This study identifies a critical vulnerability in SNNs on recently established bit-based codes: consistent performance degradation when temporal spike encoding orders are reversed, and measures the per-timestep class-mutual-information profile of six encodings directly and shows that the resulting concordance ordering predicts the observed degradation.
N. T. Luu, Trung Duong Trung Luu, N. Pham et al.· Neuromorphic Computing and E...· 0 citations
Phenomenological spiking neuron models such as Izhikevich, adaptive quadratic integrate-and-fire (aQIF), and Adaptive Exponential (AdEx) are widely used because of their simplicity and numerical efficiency. These models reproduce diverse neuronal dynamics through a slow self-inhibitory adaptation variable. Here we introduce their symmetric counterpart by replacing adaptation with slow self-excitation, motivated by intrinsic calcium-mediated membrane currents. This minimal modification enables robust persistent spiking and working-memory dynamics without compromising computational efficiency. These properties remain in excitatory spiking neural networks. We then derive and validate a mean-field neural mass model that remains stable while retaining working-memory functionality. Additionally, we implement the single-neuron model in a minimal memristor-based neuromorphic circuit and experimentally confirm its dynamics. These results provide scalable tools for large-scale brain simulations and neuromorphic applications in robotics, brain-machine interfaces, and edge AI devices.
D. Depannemaecker, Adrien D’hollande, G. Casagrande et al.· Nature Communications· 0 citations
Extensive experiments on static and neuromorphic benchmarks show that lower-bit BASC models match or outperform higher-bit baselines and retain this accuracy advantage after structured pruning, while further reducing model storage and synaptic operations.
Linliang Chen, Yan Zhong, Xin Liu et al.· 0 citations
The proposed parallel multi-compartment spiking neuron (PMSN) presents a promising solution to harness the computational advantages of detailed biological neurons, enabling high-performance and efficient temporal processing on neuromorphic computing systems.
Xinyi Chen, Jibin Wu, Chenxiang Ma et al.· IEEE Transactions on Neural...· 16 citations· ⚡4
