Entry Ma:2009:BMW from talip.bib
Last update: Sun Oct 15 02:55:04 MDT 2017
Top |
Symbols |
Numbers |
Math |
A |
B |
C |
D |
E |
F |
G |
H |
I |
J |
K |
L |
M |
N |
O |
P |
Q |
R |
S |
T |
U |
V |
W |
X |
Y |
Z
BibTeX entry
@Article{Ma:2009:BMW,
author = "Yanjun Ma and Andy Way",
title = "Bilingually Motivated Word Segmentation for
Statistical Machine Translation",
journal = j-TALIP,
volume = "8",
number = "2",
pages = "7:1--7:??",
month = may,
year = "2009",
CODEN = "????",
DOI = "https://doi.org/10.1145/1526252.1526255",
ISSN = "1530-0226 (print), 1558-3430 (electronic)",
ISSN-L = "1530-0226",
bibdate = "Wed Jun 3 16:13:52 MDT 2009",
bibsource = "http://portal.acm.org/;
http://www.math.utah.edu/pub/tex/bib/talip.bib",
abstract = "We introduce a bilingually motivated word segmentation
approach to languages where word boundaries are not
orthographically marked, with application to
Phrase-Based Statistical Machine Translation (PB-SMT).
Our approach is motivated from the insight that PB-SMT
systems can be improved by optimizing the input
representation to reduce the predictive power of
translation models. We firstly present an approach to
optimize the existing segmentation of both source and
target languages for PB-SMT and demonstrate the
effectiveness of this approach using a Chinese--English
MT task, that is, to measure the influence of the
segmentation on the performance of PB-SMT systems. We
report a 5.44\% relative increase in Bleu score and a
consistent increase according to other metrics. We then
generalize this method for Chinese word segmentation
without relying on any segmenters and show that using
our segmentation PB-SMT can achieve more consistent
state-of-the-art performance across two domains. There
are two main advantages of our approach. First of all,
it is adapted to the specific translation task at hand
by taking the corresponding source (target) language
into account. Second, this approach does not rely on
manually segmented training data so that it can be
automatically adapted for different domains.",
acknowledgement = ack-nhfb,
articleno = "7",
fjournal = "ACM Transactions on Asian Language Information
Processing",
journal-URL = "http://portal.acm.org/browse_dl.cfm?&idx=J820",
keywords = "alignment; bilingually motivated; phrase-based
statistical machine translation; word segmentation",
}
Related entries
- according,
5(2)146,
5(2)165,
6(2)7,
7(2)7,
7(4)11,
8(1)3,
9(2)6,
10(1)3,
10(4)19,
12(4)17,
13(2)9
- account,
8(2)8,
8(4)18
- achieve,
7(3)9,
7(3)10,
7(4)12,
7(4)13,
8(1)4,
8(3)10,
8(4)17,
9(2)6,
9(4)14,
10(2)7,
11(3)8,
11(3)11,
11(4)15,
11(4)17,
12(1)3,
12(1)4,
12(2)7,
12(3)9,
12(3)11,
13(1)3,
13(1)4,
13(2)9
- across,
6(3)10,
11(2)6,
11(3)11
- adapted,
11(1)2
- advantage,
9(3)11,
11(2)7,
12(3)12
- alignment,
5(2)121,
7(1)1,
7(3)10,
7(4)12,
8(2)6,
8(2)8,
11(4)18,
12(3)9,
12(4)16
- all,
6(4)2,
7(1)1,
7(2)7,
7(3)8,
7(4)11,
7(4)12,
8(4)16,
8(4)17,
9(2)5,
9(3)11,
9(3)12,
10(2)9,
10(3)15,
11(2)6,
11(2)7,
11(4)18,
12(2)5,
12(3)9,
13(1)2,
13(2)6,
13(4)16
- any,
8(4)16,
11(2)5,
12(1)4,
13(1)1
- application,
1(2)103,
4(4)475,
6(2)7,
6(4)2,
7(1)2,
7(2)7,
7(3)8,
7(4)11,
7(4)13,
8(2)9,
8(3)10,
8(4)14,
9(2)7,
9(4)13,
10(2)10,
10(4)19,
11(4)17,
12(3)11
- art, state-of-the-,
6(4)1,
6(4)3,
7(2)7,
7(4)12,
7(4)13,
8(1)4,
8(2)6,
9(2)5,
11(1)1,
11(3)8,
11(4)15,
12(1)1,
12(4)14,
12(4)16,
13(3)13
- automatically,
3(4)227,
5(2)89,
5(2)121,
5(2)165,
7(1)1,
7(2)6,
8(1)3,
8(3)10,
9(2)6,
10(3)12,
11(2)6,
11(4)16,
12(2)7,
12(4)15,
13(3)12,
13(4)16
- Based, Phrase-,
8(4)15
- based, phrase-,
8(1)4,
8(2)8,
8(4)15,
11(3)8,
12(4)16,
13(1)2
- both,
6(2)6,
6(2)7,
6(3)10,
7(1)1,
7(1)2,
7(1)3,
7(3)8,
8(3)11,
8(3)12,
9(1)2,
9(1)4,
9(2)5,
10(1)2,
10(1)4,
10(1)6,
10(2)10,
10(3)13,
10(3)15,
10(4)20,
10(4)21,
11(2)4,
11(2)6,
11(2)7,
11(3)9,
11(4)18,
12(2)5,
12(2)7,
12(4)17,
13(2)9,
13(4)17
- boundary,
7(3)10,
10(1)6,
12(1)2
- Chinese--English,
1(2)103,
4(2)57,
8(4)17,
10(4)17,
13(1)2
- consistent,
12(3)9
- corresponding,
5(2)89,
7(3)10,
9(2)7,
9(3)12,
10(2)7,
11(2)5,
11(3)8
- data,
2(2)143,
6(1)z,
6(1)z-1,
6(2)7,
6(3)11,
7(1)3,
7(3)9,
7(4)13,
8(1)3,
8(3)10,
8(3)11,
8(3)12,
8(4)16,
8(4)18,
9(2)6,
10(2)7,
10(3)12,
10(4)20,
11(2)4,
11(3)10,
11(3)11,
11(4)13,
11(4)14,
11(4)18,
12(1)1,
12(2)7,
12(3)9,
13(1)2,
13(1)3,
13(1)4,
13(4)16,
13(4)17,
13(4)18
- demonstrate,
5(2)121,
6(4)1,
7(1)2,
7(2)7,
7(3)10,
8(2)9,
9(1)4,
9(3)11,
11(2)6,
11(3)8,
13(1)1,
13(1)3
- different,
5(2)89,
6(3)9,
6(4)3,
7(2)7,
7(3)8,
7(4)13,
8(1)2,
8(2)8,
8(3)11,
8(4)16,
8(4)17,
9(1)1,
9(1)4,
9(2)5,
9(2)6,
9(3)12,
10(1)4,
10(1)5,
10(3)12,
10(4)17,
10(4)19,
11(3)8,
11(3)11,
11(4)16,
11(4)17,
11(4)18,
12(1)2,
12(3)11,
12(4)17,
13(2)6,
13(3)11,
13(4)16
- do,
7(3)8,
7(3)10,
7(4)11,
7(4)12,
7(4)13,
8(2)9,
12(1)3,
13(2)9
- domain,
5(2)165,
7(1)2,
7(4)11,
8(2)8,
10(1)5,
11(2)7,
12(1)3,
13(3)12
- effectiveness,
6(2)6,
6(2)8,
6(4)2,
7(2)5,
7(3)10,
8(4)17,
9(2)5,
9(3)11,
9(3)12,
10(3)15,
11(2)5,
11(4)18,
12(4)16,
12(4)17,
13(2)6,
13(3)13,
13(4)18
- English, Chinese-,
1(2)103,
4(2)57,
8(4)17,
10(4)17,
10(4)18,
10(4)19,
13(1)2
- existing,
6(2)8,
6(4)2,
8(1)2,
9(1)3,
9(2)5,
9(4)13,
10(3)16,
11(2)4,
11(2)7,
11(4)18,
12(2)5,
12(4)17,
13(1)2,
13(2)8
- first,
5(2)165,
6(2)6,
6(4)3,
7(1)1,
7(3)8,
7(3)10,
8(3)10,
8(3)11,
8(4)19,
9(3)10,
9(3)11,
10(3)13,
11(2)6,
11(2)7,
11(3)8,
11(3)9,
11(3)11,
12(1)3,
12(1)4,
12(2)5,
12(3)10,
12(4)17,
13(2)7,
13(2)9,
13(3)13,
13(4)17
- firstly,
10(1)6
- hand,
11(2)7
- improved,
6(3)9,
7(4)12,
8(2)6,
9(3)11,
10(1)3,
10(4)17,
11(3)10,
12(4)14,
12(4)17
- increase,
7(2)5,
7(3)9,
8(4)15,
9(3)12,
9(4)15,
10(2)8,
11(2)6,
11(4)18,
12(2)6,
12(3)11
- influence,
7(3)10,
12(3)11
- input,
5(2)165,
9(1)1,
9(1)3,
9(2)6,
10(1)6,
10(4)20,
12(1)2,
12(1)4,
12(2)6,
13(1)4,
13(2)8,
13(3)12,
13(4)17
- insight,
12(4)14
- introduce,
5(2)121,
7(4)11,
8(3)12,
10(3)16,
12(3)9,
12(4)15,
13(1)3,
13(1)4,
13(3)14
- machine,
4(1)18,
4(4)377,
5(2)89,
5(3)185,
6(2)6,
6(4)2,
6(4)3,
7(1)1,
7(2)5,
7(2)7,
7(3)9,
7(3)10,
8(1)4,
8(2)5,
8(2)6,
8(2)8,
8(2)9,
8(3)10,
8(4)15,
9(1)3,
9(4)13,
10(1)2,
10(1)5,
10(3)16,
10(4)18,
10(4)20,
11(1)2,
11(3)8,
11(4)14,
11(4)16,
12(3)9,
12(3)12,
12(4)14,
12(4)16,
12(4)17,
13(1)2,
13(3)11,
13(4)16,
13(4)17
- main,
6(3)9,
7(1)1,
7(1)2,
8(3)12,
9(3)11,
11(3)8,
12(4)16,
13(1)4
- manually,
5(2)89,
5(2)165,
9(4)14,
10(2)7,
13(3)11
- measure,
5(2)89,
6(2)6,
6(4)3,
9(2)7,
10(1)2,
10(1)6,
10(4)20,
11(2)6,
11(3)9,
11(3)11,
13(3)11,
13(3)13,
13(3)14
- metric,
7(2)7,
8(4)15,
10(1)3,
11(3)9
- more,
5(2)146,
6(2)7,
6(3)10,
7(3)9,
7(4)13,
8(1)4,
8(3)12,
8(4)14,
8(4)16,
9(1)2,
9(2)6,
9(3)11,
9(3)12,
10(1)4,
10(4)19,
11(2)4,
11(2)7,
12(1)1,
12(1)2,
12(3)9,
13(1)1,
13(1)3,
13(1)4,
13(2)8,
13(4)18
- motivated,
10(3)14,
10(4)20,
11(3)10
- MT,
2(2)143,
8(2)6,
10(3)16,
10(4)18,
12(4)16
- not,
5(2)89,
5(2)121,
6(3)10,
7(2)7,
7(3)8,
7(3)10,
7(4)11,
7(4)12,
7(4)13,
8(1)3,
8(2)8,
8(2)9,
8(4)17,
9(2)6,
9(2)7,
9(4)14,
10(1)4,
10(1)5,
10(2)8,
10(2)10,
10(4)18,
11(1)2,
11(3)10,
12(1)2,
12(1)3,
12(2)7,
12(3)9,
13(2)7,
13(2)9,
13(4)17,
13(4)18
- of-the-art, state-,
6(4)1,
6(4)3,
7(2)7,
7(4)12,
7(4)13,
8(1)4,
9(2)5,
11(1)1,
11(3)8,
11(4)15,
12(1)1,
12(4)14,
12(4)16,
13(3)13
- other,
5(2)165,
7(2)6,
7(2)7,
7(3)8,
7(4)11,
8(1)2,
8(4)14,
8(4)16,
8(4)17,
9(1)1,
10(2)7,
10(4)20,
11(2)4,
11(2)7,
11(4)18,
12(1)1,
12(1)3,
12(3)11,
12(4)16,
13(1)3,
13(1)4,
13(2)7
- performance,
5(2)121,
5(2)165,
6(2)8,
6(3)9,
6(4)1,
6(4)3,
7(1)1,
7(1)2,
7(2)5,
7(2)6,
7(2)7,
7(3)9,
7(3)10,
7(4)13,
8(1)2,
8(1)3,
8(2)8,
8(2)9,
8(3)10,
8(4)16,
8(4)17,
8(4)18,
9(1)2,
9(1)4,
9(2)5,
9(2)6,
9(3)11,
9(3)12,
9(4)14,
10(2)8,
10(3)13,
10(3)14,
11(2)7,
11(3)10,
11(3)11,
11(4)14,
11(4)15,
11(4)17,
12(1)2,
12(3)9,
12(3)11,
12(4)14,
12(4)15,
12(4)16,
13(1)3,
13(1)4,
13(2)7,
13(2)9,
13(4)16,
13(4)17
- Phrase-Based,
8(4)15
- phrase-based,
8(1)4,
8(2)8,
8(4)15,
11(3)8,
12(4)16,
13(1)2
- power,
9(4)14,
12(1)2
- predictive,
12(3)9
- present,
5(2)89,
5(2)165,
6(2)7,
6(3)10,
6(4)2,
7(1)2,
7(1)3,
7(2)7,
7(3)9,
7(4)11,
7(4)13,
8(1)3,
8(2)6,
8(2)8,
8(3)10,
8(4)14,
8(4)16,
8(4)17,
8(4)18,
8(4)19,
9(1)1,
9(1)2,
9(1)3,
9(2)6,
9(4)14,
10(1)4,
10(1)6,
10(2)7,
10(3)14,
10(4)18,
10(4)19,
11(1)2,
11(1)3,
11(2)4,
11(2)5,
11(2)6,
11(3)10,
11(4)13,
12(1)3,
12(2)5,
12(3)9,
12(3)11,
12(4)15,
13(2)8,
13(4)16
- reduce,
6(4)2,
11(3)10,
12(1)2,
12(2)5,
12(3)11,
12(3)12,
13(2)8,
13(3)14
- relative,
7(3)10,
9(3)11,
12(1)2,
12(3)9,
13(1)2
- rely,
8(2)8,
12(3)9
- relying,
11(4)18,
13(1)4
- report,
8(4)16,
9(1)3,
10(2)9,
10(2)10,
13(2)7,
13(3)11
- representation,
1(3)207,
6(2)7,
8(3)12,
8(4)14,
11(4)15,
13(2)9
- score,
7(2)7,
7(4)12,
8(3)10,
9(1)3,
9(2)6,
10(4)18,
11(2)6,
12(3)9,
12(4)17,
13(1)2,
13(3)13,
13(4)16
- second,
5(2)89,
8(1)4,
8(4)19,
11(1)3,
11(2)6,
11(3)8,
11(3)9,
11(3)11,
12(1)3,
13(2)7,
13(4)17
- segmentation,
8(4)16,
9(1)2,
9(1)3,
9(2)5,
9(4)15,
11(2)7,
12(1)2,
12(1)4,
12(4)16,
13(2)9
- segmented,
7(3)10,
8(3)11,
9(2)5,
12(1)4,
13(2)9
- segmenter,
9(2)5
- show,
5(2)89,
5(2)146,
7(1)1,
7(1)2,
7(1)3,
7(4)11,
7(4)12,
7(4)13,
8(1)4,
8(2)9,
8(3)12,
8(4)16,
8(4)17,
9(1)1,
9(1)2,
9(1)3,
9(2)5,
9(2)6,
9(2)7,
9(3)11,
9(3)12,
9(4)14,
10(1)3,
10(3)15,
11(2)4,
11(2)5,
11(2)7,
11(3)8,
11(3)11,
11(4)14,
11(4)15,
11(4)17,
11(4)18,
12(1)2,
12(1)4,
12(2)5,
12(2)7,
12(3)9,
12(3)10,
12(3)11,
12(4)15,
12(4)16,
13(1)3,
13(2)6,
13(2)7,
13(2)9,
13(3)14
- source,
5(2)121,
6(3)11,
8(2)6,
8(2)8,
8(4)17,
9(1)1,
9(4)15,
10(3)12,
10(4)17,
10(4)19,
11(4)18,
12(4)15,
13(1)1,
13(1)2,
13(3)12,
13(4)17
- specific,
7(2)7,
8(4)14,
11(4)15,
11(4)17,
12(1)3,
12(1)4,
12(2)6
- state-of-the-art,
6(4)1,
6(4)3,
7(2)7,
7(4)12,
7(4)13,
8(1)4,
8(2)6,
9(2)5,
11(1)1,
11(3)8,
11(4)15,
12(1)1,
12(4)14,
12(4)16,
13(3)13
- statistical,
1(1)3,
3(2)87,
3(4)243,
5(2)121,
5(4)323,
5(4)360,
6(1)z-4,
7(1)1,
8(1)2,
8(1)4,
8(2)6,
8(2)8,
8(2)9,
8(3)10,
8(4)15,
8(4)19,
9(2)6,
9(2)7,
9(3)11,
10(4)18,
11(2)6,
11(2)7,
11(3)8,
11(4)15,
12(1)1,
12(3)12,
12(4)14,
12(4)16,
12(4)17,
13(1)2,
13(1)3,
13(1)4,
13(4)17
- taking,
6(4)1,
8(4)18
- target,
5(2)121,
6(4)1,
7(4)13,
8(1)3,
8(2)8,
8(3)12,
8(4)17,
9(1)1,
10(3)12,
10(4)17,
11(4)17,
12(2)7,
12(3)11,
12(3)12
- task,
2(1)49,
5(2)89,
5(2)121,
6(2)7,
6(3)11,
6(4)1,
6(4)3,
7(1)1,
7(1)2,
7(2)7,
7(3)10,
7(4)13,
8(1)4,
8(4)15,
8(4)16,
9(1)4,
9(2)6,
9(3)10,
9(4)14,
9(4)15,
10(1)5,
10(3)14,
10(4)18,
10(4)20,
10(4)21,
11(1)2,
11(3)8,
11(3)11,
11(4)13,
11(4)14,
11(4)17,
11(4)18,
12(1)2,
12(1)3,
12(2)5,
12(2)7,
12(3)9,
12(4)17,
13(2)10,
13(4)17
- the-art, state-of-,
6(4)1,
6(4)3,
7(2)7,
7(4)12,
7(4)13,
8(1)4,
9(2)5,
11(1)1,
11(3)8,
11(4)15,
12(1)1,
12(4)14,
12(4)16,
13(3)13
- then,
5(2)121,
6(2)6,
7(1)1,
7(3)10,
7(4)12,
8(1)4,
8(3)10,
8(3)11,
8(3)12,
8(4)14,
9(1)1,
9(2)7,
9(3)11,
10(2)7,
10(3)13,
10(4)20,
11(1)3,
11(2)7,
11(3)11,
11(4)15,
12(1)3,
12(3)10,
12(4)17,
13(1)4,
13(2)9,
13(3)13,
13(4)16
- there,
7(3)9,
7(4)11,
8(3)12,
8(4)17,
9(2)5,
9(3)12,
9(4)15,
10(1)2,
10(3)14,
12(1)2,
13(1)1,
13(2)8,
13(3)14
- training,
5(2)121,
6(3)11,
7(1)1,
8(1)3,
8(2)6,
8(3)10,
9(2)5,
10(3)12,
10(3)13,
11(3)9,
12(1)1,
12(2)5,
12(3)9,
12(3)10,
13(1)2,
13(1)3,
13(1)4,
13(4)17
- two,
5(2)89,
7(2)7,
7(3)8,
7(4)11,
7(4)12,
7(4)13,
8(1)4,
8(4)17,
9(1)2,
9(3)11,
9(4)13,
10(1)2,
10(3)12,
10(3)14,
10(3)15,
10(4)20,
11(2)4,
11(2)5,
11(2)7,
11(3)8,
11(3)9,
11(3)11,
11(4)17,
12(1)1,
12(1)2,
12(1)4,
12(2)5,
12(3)10,
12(3)11,
12(4)16,
13(1)1,
13(1)3,
13(1)4,
13(2)6,
13(2)9,
13(3)11,
13(4)17
- Way, Andy,
10(4)17
- where,
7(3)8,
7(3)9,
7(3)10,
8(2)8,
9(1)3,
9(2)7,
9(4)13,
10(4)18,
11(2)4,
11(4)16,
12(1)3,
12(4)14