p22 / Journal
Evaluation of bio-inspired models under different learning settings for energy efficiency in network traffic prediction
International Journal of Neural Systems 35(12), 2550058 · 2025
In plain language
Can biologically inspired neural models forecast cellular traffic while consuming less energy?
Research context
The study encourages network forecasting systems to measure resource costs alongside accuracy.
Approach
Spiking neural networks and echo state networks are compared with CNNs and MLPs in centralized and federated settings, using traffic from three Barcelona locations.
Findings
The evaluated bio-inspired models reduce energy use while retaining prediction quality comparable to conventional architectures.
Summary based on abstract.