Abstract:We introduce a simple time-triggered protocol to achieve communication-efficient non-Bayesian learning over a network. Specifically, we consider a scenario where a group of agents interact over a graph with the aim of discerning the true state of the world that generates their joint observation profiles. To address this problem, we propose a novel distributed learning rule wherein agents aggregate neighboring beliefs based on a min-protocol, and the inter-communication intervals grow geometrically at a rate $a \geq 1$. Despite such sparse communication, we show that each agent is still able to rule out every false hypothesis exponentially fast with probability $1$, as long as $a$ is finite. For the special case when communication occurs at every time-step, i.e., when $a=1$, we prove that the asymptotic learning rates resulting from our algorithm are network-structure independent, and a strict improvement upon those existing in the literature. In contrast, when $a>1$, our analysis reveals that the asymptotic learning rates vary across agents, and exhibit a non-trivial dependence on the network topology coupled with the relative entropies of the agents' likelihood models. This motivates us to consider the problem of allocating signal structures to agents to maximize appropriate performance metrics. In certain special cases, we show that the eccentricity centrality and the decay centrality of the underlying graph help identify optimal allocations; for more general scenarios, we bound the deviation from the optimal allocation as a function of the parameter $a$, and the diameter of the communication graph.

A Fast Bayesian Iterative Rule in Amoeba Algorithm.

Rapid Physarum Algorithm for Shortest Path Problem

Bayesian Analysis for Exponential Random Graph Models Using the Adaptive Exchange Sampler.

A Communication-Efficient Algorithm for Exponentially Fast Non-Bayesian Learning in Networks

A modified pulse coupled neural network for shortest-path problem

A neural network method for the computing of the shortest path tree

Modeling Nonlinear Dynamic Biological Systems with Human-Readable Fuzzy Rules Optimized by Convergent Heterogeneous Particle Swarm

Amoeba Monte Carlo algorithms for random trees with controlled branching activity: efficient trial move generation and universal dynamics

Fast Algorithms for Simulation of Neuronal Dynamics Based on the Bilinear Dendritic Integration Rule

A Fusion Algorithm Based on Physarum Polycephalum Network and Ant Colony Optimization Algorithm

Network Division Method Based on Cellular Growth and Physarum-inspired Network Adaptation

Applications to Biological Networks of Adaptive Hagen-Poiseuille Flow on Graphs

Distributed Algorithms from Arboreal Ants for the Shortest Path Problem

A biologically inspired network design model

Minimize the Average Mean First Passage Time of Random Walk in Complex Networks by Genetic Algorithm.

Local Message Passing on Frustrated Systems

High Performance Computing of Gene Regulatory Networks using a Message-Passing Model

Energy efficient permanence‐based community detection algorithm

When does the Physarum Solver Distinguish the Shortest Path from other Paths: the Transition Point and its Applications

P-bRS: A Physarum-Based Routing Scheme for Wireless Sensor Networks

PANDA: A physarum-inspired algorithm to solve the multi-objective discrete network design problem