Fast exploration and learning of latent graphs with aliased observations

TL;DR

This paper introduces an efficient algorithm for exploring and recovering latent graph structures with aliased observations, significantly outperforming naive methods in complex topologies while maintaining competitiveness in unaliased scenarios.

Contribution

The paper presents a novel, exponentially faster exploration algorithm for latent graphs with aliased observations, advancing the understanding of POMDP recovery.

Findings

Exponential speedup over naive exploration in challenging topologies

Competitive performance with existing methods in unaliased regimes

Effective recovery of latent graph topology with minimal steps

Abstract

We consider the problem of recovering a latent graph where the observations at each node are \emph{aliased}, and transitions are stochastic. Observations are gathered by an agent traversing the graph. Aliasing means that multiple nodes emit the same observation, so the agent can not know in which node it is located. The agent needs to uncover the hidden topology as accurately as possible and in as few steps as possible. This is equivalent to efficient recovery of the transition probabilities of a partially observable Markov decision process (POMDP) in which the observation probabilities are known. An algorithm for efficiently exploring (and ultimately recovering) the latent graph is provided. Our approach is exponentially faster than naive exploration in a variety of challenging topologies with aliased observations while remaining competitive with existing baselines in the unaliased regime.