# M q t prof je bilmes m q 1 pt qt q x pt

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

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ¯t |Qt = q ) max p(Qt = q |Qt−1 = x r (5.15) m r)αr (t − 1) (5.16) We get the ﬁnal max can be computed from the ﬁnal max marginal: Prof. Jeﬀ Bilmes EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-41 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation We can thus deﬁne a modiﬁed form of the α-recursion that, rather than uses summation, uses a max operator. m αq (t) m αq (1) = p(¯t |Qt = q ) x = p(¯t |Qt = q ) max p(Qt = q |Qt−1 = x r (5.15) m r)αr (t − 1) (5.16) We get the ﬁnal max can be computed from the ﬁnal max marginal: max q1:T ∈DQ1:T Prof. Jeﬀ Bilmes m p(¯1:T , q1:T ) = max αq (T ) x (5.17) q EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-41 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation We can thus deﬁne a modiﬁed form of the α-recursion that, rather than uses summation, uses a max operator. m αq (t) m αq (1) = p(¯t |Qt = q ) x = p(¯t |Qt = q ) max p(Qt = q |Qt−1 = x r (5.15) m r)αr (t − 1) (5.16) We get the ﬁnal max can be computed from the ﬁnal max marginal: max q1:T ∈DQ1:T m p(¯1:T , q1:T ) = max αq (T ) x (5.17) q Max operator is similar to sum (a commutative semi-ring), in that it marginalizes out hidden variables. Prof. Jeﬀ Bilmes EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-41 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation Max operator is similar to sum, in that it marginalizes out hidden variables. Prof. Jeﬀ Bilmes EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-42 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation Max operator is similar to sum, in that it marginalizes out hidden variables. Given p(¯, q ) when we form maxq p(¯, q ) we can think of this as a x x marginal, the “max marginal” of the form pm (¯) = max p(¯, q ) x x (5.18) q Prof. Jeﬀ Bilmes EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-42 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation Max operator is similar to sum, in that it marginalizes out hidden variables. Given p(¯, q ) when we form maxq p(¯, q ) we can think of this as a x x marginal, the “max marginal” of the form pm (¯) = max p(¯, q ) x x (5.18) q m Given this, we can view the αq (t) as the max marginals up to time t m αq (t) = pm (x1:t , Qt = q ) (5.19) so that the above ﬁnal maximization makes sense. We’ve deﬁned a recursive way to compute the max marginal. Prof. Jeﬀ Bilmes EE596A/Winter 2013/DGMs – Lecture 5 - Jan 25th, 2013 page 5-42 (of 232) HMMs Trellis Other HMM queries MPE Sampling What HMMs can do Summary Scratch Most Probable Explanation Max operator is similar to sum, in that it marginalizes out hidden variables. Given p(¯, q ) when we form maxq p(¯, q ) we can think of th...
View Full Document

## This document was uploaded on 04/05/2014.

Ask a homework question - tutors are online