Uncertainty Estimation in Multi-Agent Distributed Learning