Blackboard Mixture of Experts: A Unique Edge Architecture for Multi-Agent Neural Networks
Oliverio Colon · Zenodo (CERN European Organization for Nuclear Research) · 2026
The paper introduces a resource-efficient edge architecture designed to bring advanced multi-agent capabilities and spatial awareness to resource-constrained microcontrollers (e.g., ARM Cortex-M, RP2040) without the heavy compute and memory overhead of traditional floating-point models.