
The asymptotic mutual information between an unknown signal belonging to a high-dimensional compact group and its noisy pairwise observations can be approximated by the mutual information in a simpler linear observation model. This result enables the characterization of the information-theoretic limits of estimation in group synchronization and quadratic assignment problems.


coremsg

asymptotic-mutual-information-and-minimum-mean-squared-error-in-bayesian-inference-of-signals-over-compact-groups


Asymptotic Mutual Information and Minimum Mean-Squared Error in Bayesian Inference of Signals over Compact Groups
