SpikeVLA: Vision-Language-Action Models with Spiking Neural Networks

2026-06-26

AI Summary

This work addresses the high inference latency and energy consumption of existing vision–language–action (VLA) models, which rely on large Transformers and are thus ill-suited for deployment in low-power, real-time embodied intelligence scenarios. To overcome these limitations, the study introduces spiking neural networks into the VLA framework for the first time, proposing an efficient architecture comprising a spiking vision encoder (Spike-V), a multimodal spiking large language model (Spike-L), and a spiking action policy network (Spike-A) based on Laplacian kernel population coding. By leveraging event-driven sparse computation and spiking dynamics, the proposed method substantially reduces both energy consumption and computational overhead while maintaining competitive task performance in navigation and robotic control tasks, thereby establishing a new paradigm for low-power, real-time embodied AI.

Abstract

Vision-Language-Action (VLA) models have become a dominant paradigm for embodied intelligence. However, most existing approaches are built on large-scale transformers, resulting in substantial inference latency and energy consumption that limit their practical deployment in low-power, real-time scenarios. We propose SpikeVLA, a spiking VLA architecture for embodied navigation with energy-efficient inference, consisting of three key components. (i) A spiking vision encoder, Spike-V, that replaces dense continuous layers with event-driven spiking layers to reduce the energy consumption of visual representation learning. (ii) A multi-modal spiking large language model, Spike-L, that reformulates cross-modal reasoning with spiking dynamics and token-level event-driven sparsity to further lower computational cost. (iii) A spiking action policy network, Spike-A employs Laplacian-kernel population coding with a multi-layer fully connected SNN, and decodes spiking activities into stable and robust continuous control with energy-efficient inference under low-power constraints. Experiments on navigation and robotic control tasks show that SpikeVLA significantly reduces energy consumption and computational cost while maintaining competitive performance, highlighting its potential for low-power, real-time embodied intelligence.

Problem

• Vision-Language-Action models
• energy efficiency
• spiking neural networks
• real-time inference
• embodied intelligence

Innovation

• Spiking Neural Networks
• Vision-Language-Action Models
• Energy-Efficient Inference
• Event-Driven Sparsity
• Embodied Intelligence