Several people have contributed to precursors to this library
or contributed key components in the ASR system used to
develop this library.
Emerald Chung wrote the early versions of the recognizer.
Donald Hindle provided us with the n-gram language
models used in our experiments.
Andrej Ljolje provided us with the acoustic models for
our experiments and early versions of the acoustic model likelihood
computation.
Fernando Pereira was an FSM library author
and contributed to its ASR application,
particularly in the recognition transducer
construction and in lattice generation.
Murat Saraclar wrote the HTK
dynamic-shared-object acoustic model module
and the HTK context-dependency transducer constructor
available for use with this recognizer.