"Picking a role model is a natural way of deciding what to do by emulating a person we admire: children look up to superheroes; chemists and physicists follow Marie Curie. When faced with a difficult choice, solving the problem on our own may seem difficult. Instead, we seek shelter behind our role model and try to imagine what Superman or Marie Curie would have done. There is a parallel to this approach in statistical signal processing. Consider the design of an estimator based on low quality observations. Bayesian theory dictates rules for the optimal design of such an estimator. Suppose there exists a better estimator that has access to superior observations. We can use this better estimator as a role model, and train our estimator to imitate its outputs. Under certain conditions, we have proved that this role-model approach gives the same optimal solution as the Bayesian approach, albeit with a different design methodology that is easier to implement. The motivation for this approach is in adaptive receiver design for wireless communications. During the proposed fellowship, we plan to extend the theory for the role-model framework, to refine its application to receiver design, to study its connection with other problems in information theory, and to investigate possible interdisciplinary applications. For wireless communications, the project will result in optimized low-complexity receiver structures that can save power, increase radio coverage and data integrity, and reduce the cost of mobile handsets. Further applications of role models may include computer science, risk assessment, image processing, and others."