Speech Recognition Algorithms based on Weighted Finite-State Transducers (Synthesis Lectures on Speech and Audio Processing)