Storing the max scores okt overall ot n 2 kt prof

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: 0) k-best without the penalty Island Summary Scratch Compute Cost max marginal construction O(T N 2 ) block construction O(kT N ). computing max scores: O(kT log(kT )) Dynamic programming: O(kT N ). Overall cost: overall computational costs of finding the k -best assignments is O(T N 2 + kT log(kT ) + kT N ). Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 9 - Feb 6th, 2013 page 9-27 (of 180) k-best without the penalty Island Summary Scratch Memory Cost Storing max marginals: O(T N 2 ) Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 9 - Feb 6th, 2013 page 9-28 (of 180) k-best without the penalty Island Summary Scratch Memory Cost Storing max marginals: O(T N 2 ) storing k -best paths: O(T k ). Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 9 - Feb 6th, 2013 page 9-28 (of 180) k-best without the penalty Island Summary Scratch Memory Cost Storing max marginals: O(T N 2 ) storing k -best paths: O(T k ). storing the max scores: O(kT ) Prof. Jeff Bilmes EE596A/Winter 2013/DGMs – Lecture 9 - Feb 6th, 2013 page 9-28 (of 180) k-best without the penalty Island Summary Scratch Memory Cost Storing max marginals: O(T N 2 ) storing k -best paths: O(T k ). storin...
View Full Document

Ask a homework question - tutors are online