Tag
SpikingVLA introduces an ANN-to-SNN conversion framework using Dendritic Integrate-and-Fire neurons and asynchronous execution for energy-efficient, low-latency spiking VLA inference, improving navigation success and reducing first-action latency by 11.2x over existing spiking VLA methods.
This paper proposes a plug-and-play framework that implements spike-friendly approximations for Transformer nonlinearities (e.g., Softmax, SiLU, normalization) via population computation with LIF neurons and lightweight bit-shift scaling, achieving less than 1% accuracy drop on LLMs without fine-tuning.