State-Space Model Inspired Multiple-Input Multiple-Output Spiking Neurons

2025-04-03

AI Summary

To address the limited representational capacity of conventional single-input single-output (SISO) neuron models in spiking neural networks (SNNs), this paper proposes the first state-space-based multi-input multi-output (MIMO) spiking neuron model. The neuron is formulated as a dynamic system featuring linear state evolution and nonlinear spike generation, supporting flexible configurations—including single-input multiple-output (SIMO), multiple-input single-output (MISO), and full MIMO. Crucially, it introduces high-dimensional internal states coupled with a multi-channel output coordination mechanism, substantially enhancing representational capacity under constrained neuron counts. Experiments demonstrate that, at equivalent network scale, SNNs employing the proposed MIMO neurons achieve accuracy comparable to continuous-valued baselines—validating that increasing per-neuron output dimensionality effectively boosts performance in resource-constrained SNNs. This work establishes a novel architectural paradigm for SNN design grounded in state-space modeling.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

In spiking neural networks (SNNs), the main unit of information processing is the neuron with an internal state. The internal state generates an output spike based on its component associated with the membrane potential. This spike is then communicated to other neurons in the network. Here, we propose a general multiple-input multiple-output (MIMO) spiking neuron model that goes beyond this traditional single-input single-output (SISO) model in the SNN literature. Our proposed framework is based on interpreting the neurons as state-space models (SSMs) with linear state evolutions and non-linear spiking activation functions. We illustrate the trade-offs among various parameters of the proposed SSM-inspired neuron model, such as the number of hidden neuron states, the number of input and output channels, including single-input multiple-output (SIMO) and multiple-input single-output (MISO) models. We show that for SNNs with a small number of neurons with large internal state spaces, significant performance gains may be obtained by increasing the number of output channels of a neuron. In particular, a network with spiking neurons with multiple-output channels may achieve the same level of accuracy with the baseline with the continuous-valued communications on the same reference network architecture.

Problem

• Extends SISO to MIMO spiking neuron models
• Explores SSM-based neurons with linear/non-linear dynamics
• Demonstrates multi-output channels boost SNN performance

Innovation

• MIMO spiking neuron model based on SSMs
• Linear state evolutions with non-linear activations
• Multiple-output channels enhance performance significantly