Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
37987 | World Patent Information | 2013 | 11 Pages |
Abstract
This article provides a guide for readers interested in deepening their knowledge of the strengths and weaknesses of the different machine translation (MT) quality metrics, and presents a methodology and tooling developed respectively by LexisNexis® and its MT provider Asia Online™ as part of a human quality assessment framework for patent translation. The methodology is designed specifically to compensate for the shortcomings of automated evaluation.
► Assessment of different MT quality metrics. ► Problems and approaches to this in the patent field. ► Merits and challenges of automated evaluation. ► Approach using human evaluation of MT explained and tested.
Keywords
Related Topics
Physical Sciences and Engineering
Chemical Engineering
Bioengineering
Authors
Laura Rossi, Dion Wiggins,