77 search results - page 7 / 16 » Language Models for Machine Translation: Original vs. Transl... |
EMNLP
2010
13 years 4 months ago
2010
Minimum Error Rate Training is the algorithm for log-linear model parameter training most used in state-of-the-art Statistical Machine Translation systems. In its original formula...
COLING
2000
13 years 8 months ago
2000
When aligning texts in very different languages such as Korean and English, structural features beyond word or phrase give useful intbrmation. In this paper, we present a method f...
ICMLC
2010
Springer
13 years 4 months ago
2010
Springer
: In this paper, we propose a new approach to improve the translation quality by adding the Key-Words of a sentence to the parallel corpus. The main idea of the approach is to find...
LREC
2010
13 years 8 months ago
2010
In Chinese texts, words composed of single or multiple characters are not separated by spaces, unlike most western languages. Therefore Chinese word segmentation is considered an ...
EMNLP
2011
12 years 6 months ago
2011
We propose a general method to watermark and probabilistically identify the structured outputs of machine learning algorithms. Our method is robust to local editing operations and...