My question is regarding the application of Maximum Likelihood (ML) bound in the context of digital communication. This bound is useful in evaluating the performance of detection algorithms.
Can someone please explain what is meant by the terms "ML Upper Bound". I read https://www.quora.com/In-algorithms-what-is-the-upper-and-lower-bound that
- Upper bound implies "the best an algorithm can do"
- Lower bound implies "There is no algorithm that can do better than this"
I always thought of ML bound as the lower bound on the performance. But saw some authors comparing the performance of their algorithms with the ML upper bound. What is the significance of such a comparison?
Any help will be much appreciated.