Selection and Combination of Hypotheses for Dialectal Speech Recognition

   Abstract