Figure 6

Abstraction augmented Markov models. (a) An abstraction hierarchy T on a set S = {s₁,…,s₉} of 2-grams over an alphabet of size 3. The abstractions a₁ to a₉ correspond to the 2-grams s₁ to s₉, respectively. The subset of nodes A = {a₁₅, a₆, a₁₄} represents a 3-cut γ₃ through T; (b) Dependency of X_i on A_i, which takes values in a set of abstractions A corresponding to an m-cut γ_m, in a k_th order AAMM.