Compressed sensing and observability of agents

Analysis

Regularized inverse problem methods (total variation, least squares) allow reconstruction from sparse, noisy measurements, which is the same objective as inferring an internal agent state from limited logs.

1 source videos