Hidden Markov Models (1989)
In the late 1960s L. E. Baum discovered a method that, given
a long sequence of observations, could identify the underlying
model that generated the sequence.
In the 1970s, based on Baum’s work, James Baker created a
speech recognition system that outperformed the current systems. Unlike most systems, his relied on no linguistic knowledge, instead only advanced statistics, to do pattern recognition
of speech signals.
Baker’s approach intrigued Lee Neuwirth, the director of the
Institute of Defense Analysis, who named the models “Hidden
Markov Models.”
Lawrence R. Rabiner’s tutorial in IEEE popularized HMMs.
HMMs used in:
- pattern recognition
- sequence analysis of genes
- speech recognition
- handwriting recognition
back