Fu-Hua Liu, Pedro Moreno, Richard Stern, and Alex Acero March 1994
In Proc. of ARPA Human Language Technology Workshop
