Exercise must there be a unqiue stationary

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ionary distribution, this covariance goes to zero exponentially fast. Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 4 - Jan 23rd, 2013 page 4-68 (of 239) HMMs HMMs as GMs Other HMM queries What HMMs can do MPE Summ Correlation over time of simple HMM Thus, cov(Xt , Xt+h ) is in general not equal to zero. h But recall, Ah − 1π from lecture 3, and this is a matrix with all → rows equal to the stationary distribution. Therefore, we have cov(Xt , Xt+h ) (4.70) µi µj (Ah )ij πi − = µi π i h − → µi µj πj πi − ij µi π i i ij µi π i i (4.71) i µi π i =0 (4.72) i Thus, while the covariance between to observations is not necessarily zero in an HMM, once we are at a stationary distribution, this covariance goes to zero exponentially fast. Exercise: must there be a unqiue stationary distribution here? Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 4 - Jan 23rd, 2013 page 4-68 (of 239) HMMs HMMs as GMs Other HMM queries What HMMs can do MPE Summ Correlation over time of simple HMM Example of the decay in the mutual-information correlation from a real-world HMM. I.e., we see f (τ ) = I (Xt ; Xt+τ ), where I () is the mutual information function. Mutual information is stronger than correlation. This is compared against i.i.d. samples (high peak at τ = 0 is expected). −3 14 x 10 HMM MI IID MI 13 12 MI (bits) 11 10 9 8 7 6 5 −100 Prof. Jeff Bilmes −50 0 time (ms) 50 100 EE596A/Winter 2013/DGMs – Lecture 4 - Jan 23rd, 2013 page 4-69 (of 239) HMMs HMMs as GMs Other HMM queries What HMMs can do MPE Summ State Duration Modeling Markov chain, state duration distribution is geometric Let D be such a random variable, then P (D = d) = pd−1 (1 − p) (4.73) where d ≥ 1 is an integer and p = aii , if D is the duration random variable for state i of the chain, giving: 0.25 0.2 0.75 0.15 P(d) 0.25 0.1 0.05 0 10 20 30 d 40 50 60 Many sequential tasks have sub-segments that do not follow this distribution Prof. Jeff Bilmes page 4-70 (of 239) this seems like EE596A/Winter 2013/DGMs – Lecture 4 - Jan 23rd, 2013 an inherent limitation. HMMs HMMs as GMs Other HMM queries What HMMs can do MPE Summ State Duration Modeling Many “tricks” to using an HMM, can alleviate such problems. “state-tying”, where multiple states have the same observation distribution (parameters are shared). That is, state q and q are tied if it is the case that p(x|Qt = q ) = p(x|Qt = q ) ∀x ∈ DX (4.74) If n states in a series are strung together, all of which share the same observation distribution, that observation distribution will be active for as long as we are in that state. For example: 0.08 0.75 0.75 0.25 0.75 0.25 0.06 0.75 0.25 0.25 0.04 0.02 0 Prof. Jeff Bilmes 10 20 EE596A/Winter 2013/DGMs – Lecture 4 - Jan 23rd, 2013 30 d 40 50 page 60 4-71 (of 239) HMMs HMMs as GMs Other HMM queries What HMMs can do MPE Summ State Duration Modeling This corresponds to the sum of random variables. Let {Di }i be a collection of independent geometrically distributed random variables with parameter p, and let Wr = r=1 Dr , then i k−1 r p (1 − p)k−r , k = r, r + 1, . . . r−1 p(Wr = k ) = (4.75) This is a “negati...
View Full Document

This document was uploaded on 04/05/2014.

Ask a homework question - tutors are online