Towards an Expressivity-Normalized Energy-Demand Comparison of ANNs and SNNs
MCML Authors
Miriam Kranzlmüller
Gitta Kutyniok
Prof. Dr.
Core PI
Abstract
Miriam Kranzlmüller
Gitta Kutyniok
Prof. Dr.
Core PI
Abstract
Spiking neural networks (SNNs) are often regarded as energy-efficient alternatives to artificial neural networks (ANNs), yet their advantage depends critically on both network architecture and data properties. We develop an analytical framework to compare fully-connected ReLU ANNs and integrate-and-fire SNNs for time-series data with respect to their theoretical energy efficiency at matched expressive capacity. By relating an inference-energy model to theoretical bounds on representational expressivity, we derive an expressivity-normalized efficiency ratio and explicit thresholds in network width, spike sparsity, and ANN depth scaling. Our analysis characterizes the regimes in which event-driven computation offsets the temporal overhead of SNNs, providing capacity-aware principles for designing energy-efficient temporal networks. It shows that ANNs exceed SNNs in expressivity-normalized efficiency only in specific regimes.
misc KEK26
Preprint
Aug. 2026Authors
M. Kranzlmüller • P. Esser • G. KutyniokLinks
arXivResearch Area
BibTeXKey: KEK26