New Approaches to Distributed State Estimation, Inference and Learning with Extensions to Byzantine-Resilience