Hidden Markov Model
•Observations, v(1), v(2) …, v(M).
–We never know the state, but at each time step a state produces an observation.
•Observation distribution:
–b(j,k) = P(v(k) at t| q(t) = Sj).
–Note this is also taken to be time independent.
•Example, HMM that generates contours varying from smooth to rough.