A Flexible Example Annotation Schema: Translation Corresponding 
Tree Representation 
Fai WONG, Dong Cheng HU, Yu Hang MAO 
Speech and Language Processing Research Center,  
Tsinghua University, 100084 Beijing 
huangh01@mails.tsinghua.edu.cn 
{hudc, myh-dau}@mail.tsinghua.edu.cn
Ming Chui DONG 
Faculty of Science and Technology 
of University of Macao, 
PO Box 3001, Macao SAR 
dmc@inesc-macau.org.mo 
Abstract 
This paper presents work on the task of con-
structing an example base
1
 from a given bi-
lingual corpus based on the annotation 
schema of Translation Corresponding Tree 
(TCT). Each TCT describes a translation ex-
ample (a pair of bilingual sentences). It repre-
sents the syntactic structure of source 
language sentence, and more importantly is 
the facility to specify the correspondences be-
tween string (both the source and target sen-
tences) and the representation tree. 
Furthermore, syntax transformation clues are 
also encapsulated at each node in the TCT 
representation to capture the differentiation of 
grammatical structure between the source and 
target languages. With this annotation 
schema, translation examples are effectively 
represented and organized in the bilingual 
knowledge database that we need for the Por-
tuguese to Chinese machine translation sys-
tem. 
1 Introduction 
The construction of bilingual knowledge base, in 
the development of example-based machine 
translation systems (Sato and Nagao, 1990), is 
vitally critical. In the translation process, the ap-
plication of bilingual examples concerns with 
how examples are used to facilitate translation, 
which involves the factorization of an input sen-
tence into the format of stored examples and the 
conversion of source texts into target texts in 
terms of the existing translations by referencing 
to the bilingual knowledge base. Theoretically 
speaking, examples can be achieved from bilin-
                                                           
1
 Or bilingual knowledge base, we use the two terms inter-
changeably. 
gual corpus where the texts are aligned in senten-
tial level, and technically, we need an example 
base for convenient storage and retrieval of ex-
amples. The way of how the translation examples 
themselves are actually stored is closely related 
to the problem of searching for matches. In struc-
tural example-based machine translation systems 
(Grishman, 1994; Meyers et al., 1998; Watanabe 
et al., 2000), examples in the knowledge base are 
normally annotated with their constituency (Kaji 
et al., 1992) or dependency structures (Matsu-
moto et al., 1993; Aramaki et al., 2001; Al-
Adhaileh et al., 2002), which allows the corre-
sponding relations between source and target sen-
tences to be established at the structural level. All 
of these approaches annotate examples by mean 
of a pair of analyzed structures, one for each lan-
guage sentence, where the correspondences be-
tween inter levels of source and target structures 
are explicitly linked. However, we found that 
these approaches require the bilingual examples 
that have ‘parallel’ translations or ‘close’ syntac-
tic structures (Grishman, 1994), where the source 
sentence and target sentences have explicit corre-
spondences in the sentences-pair. For example, in 
(Wu, 1995), the translation examples used for 
building the translation alignments are selected 
based on strict constraints. As a result, these ap-
proaches indirectly limit their application in us-
ing the translation examples that are ‘free 
translation’ to the development of example-
based machine translation system. In practice, 
most of the existing bilingual corpus, the mean-
ings of the source sentences are interpreted in 
target language in the nature of ‘freer’, other than 
literally translated in a projective manner and 
stayed as close to the source text as possible, in 
particular for the languages-pair that are struc-
tural divergences, such as Portuguese and Chi-
nese. 
As illustrated in Figure 1, the translation of the 
Portuguese sentence “Onde ficam as barracas de 
praia?” is interpreted into “更衣室在哪裡? 
(Where are the bathhouses?)” other than 
straightly translated to “沙灘帳篷在哪裡 ? 
(Where are the tents of beach?)”. The translations 
of the words, i.e. “barracas” and “praia”, of the 
source sentence do not explicitly appear in target 
sentence. As a result, in the conventional align-
ment process, to achieve a fully aligned structural 
representation for such sentences-pair may be 
problematic. However, we found that such type 
of examples is very common. We have investi-
gated around 2100 bilingual examples that are 
extracted from a grammar book “Gramática da 
Língua Portuguesa” (Wang and Lu, 1999), and 
found that 63.4% of examples belong to the dis-
cussed case, where the number of unmatched 
words is more than half the number of words in 
source sentence. In this paper, we overcome the 
problem by designing a flexible representation 
schema, called Translation Corresponding Tree 
(TCT). We use the TCT as the basic structure to 
annotate the examples in our example bilingual 
knowledge base for the Portuguese to Chinese 
example-based machine translation system. 
在?
Onde ficam as barracas de praia ?
哪裡更衣室
?
 
Figure 1. An example of ‘free translation’, where 
the translations of some words in Portuguese sen-
tence do not appear in target Chinese sentence. 
2 Translation Corresponding Tree 
(TCT) Representation 
TCT structure, as an extension of structure string-
tree correspondence representation (Boitet and 
Zaharin, 1988), is a general structure that can 
flexibly associate not only the string of a sen-
tence to its syntactic structure in source language, 
but also allow the language annotator to explic-
itly associate the string from its translation in 
target language for the purpose to describe the 
correspondences between different languages.  
2.1 The TCT Structure 
The TCT representation uses a triple sequence 
intervals [SNODE(n)/STREE(n)/STC(n)] en-
coded for each node in the tree to represent the 
corresponding relations between the structure of 
source sentence and the substrings from both the 
source and target sentences. In TCT structure, the 
correspondence is made up of three interrelated 
correspondences: 1) one between the node and 
the substring of source sentence encoded by the 
interval SNODE(n), which denotes the interval 
containing the substring corresponding to the 
node; 2) one between the subtree and the sub-
string of source sentence represented by the in-
terval STREE(n), which indicates the interval of 
substring that is dominated by the subtree with 
the node as root; and 3) the other between the 
subtree of source sentence and the substring of 
target sentence expressed by the interval STC(n), 
which indicates the interval containing the sub-
string in target sentence corresponding to the 
subtree of source sentence. The associated sub-
strings may be discontinuous in all cases. This 
annotation schema is quite suitable for represent-
ing translation example, where it preserves the 
strength in describing non-standard and non-
projective linguistic phenomena for a language 
(Boitet and Zaharin, 1988; Al-Adhaileh et al., 
2002), on the other hand, it allows the annotator 
to flexibly define the corresponding translation 
substring from the target sentence to the repre-
sentation tree of source sentence when it is nec-
essary. This is actually the central idea behind the 
formalism of TCT. 
NP(4/3-6/1-3)
Onde
1
ficam
2
as
3
de
5
praia
6
PP(5/5-6/Ø)
Adv(1/1/5-6) V(2/2/4)
S(2/1-6/1-6)
VP(2/2-6/1-4)
NP(4/3-4/Ø)
Syntactic Tree
Source
String
{
Det(3/3/Ø) Prep(5/5/Ø) N(6/6/Ø)
barracas
4
N(4/4/Ø)
在
4
更
1
衣
2
室
3
哪
5
裡
6
Target String{
 
Figure 2. An TCT representation for annotating 
the translation example "Onde ficam as barracas 
de praia? (Where are the bathhouses?) / 
更衣室在哪裡?" and its phrase structure together 
with the correspondences between the substrings 
(of both the source and target sentences) and the 
subtrees of sentence in source language. 
 
As illustrated in Figure 2, the translation ex-
ample “Onde ficam as barracas de praia?/ 
更衣室在哪裡?” is annotated  in a TCT struc-
ture. Based on the interpretation structure of the 
source sentence “Onde ficam as barracas de 
praia?”, the correspondences between the sub-
strings (of source and target sentences) and the 
grammatical units at different inter levels of the 
syntactic tree of the source sentence are ex-
pressed in terms of sequence intervals. The words 
of the sentences pair are assigned with their posi-
tions respectively, i.e. “Onde (1)”, “ficam (2)”, 
“as (3)”, “barracas (4)”, “de (5)” and “praia (6)” 
for the source sentence, as well as for the target 
sentence. But considering that Chinese uses 
ideograms in writing without any explicit word 
delimiters, the process to identify the boundaries 
of words is considered to be the task of word 
segmentation (Teahan et al., 2000), instead of 
assigning indices in word level with the help of 
word segmentation utility, a position interval is 
assigned to each character for the target (Chi-
nese) sentence, i.e. “更 (1)”, “衣 (2)”, “室 (3)”, 
“在 (4)”, “哪 (5)” and “裡 (6)”. Hence, a sub-
string in source sentence that corresponds to the 
node of its representation is denoted by the inter-
vals encoded in SNODE(n) for the node, e.g. the 
shaded node, NP, with interval, SNODE(NP)=4, 
corresponds to the substring “barracas” in source 
sentence that has the same interval. A substring 
of source sentence that corresponds to a subtree 
of its syntactic tree is denoted by the interval re-
corded in STREE(n) attached to the root of the 
subtree, e.g. the subtree of the shaded node, NP, 
encoded with the interval, STREE(NP)=3-6, cor-
responds to the substring “as barracas de praia” 
in source sentence. While the translation corre-
spondence between the subtree of source sen-
tence and substring in the target sentence is 
denoted by the interval assigned to the STC(n) of 
each node, e.g. the subtree rooted at shaded node, 
NP, with interval, STC(NP)=1-3, corresponds to 
the translation fragment (substring) “更衣室” in 
target sentence. 
2.2 Expressiveness of Linguistic Infor-
mation 
Another inherited characteristic of TCT structure 
is that it can be flexibly extended to keep various 
kinds of linguistic information, if they are con-
sidered useful for specific purpose, in particularly 
the linguistic information that differentiating the 
characteristics of two languages which are struc-
tural divergences (Wong et al., 2001). Basically, 
each node representing a grammatical constituent 
in the TCT annotation is tagged with grammati-
cal category (part of speech). Such feature is 
quite suitable for the describing specific linguis-
tic phenomena due to the characteristic of a lan-
guage. For instance, in our case, the crossing 
dependencies (syntax transformation rules) for 
the sentence constituents between Portuguese and 
Chinese are captured and attached to each node 
in the TCT structure for a constituent that indi-
cates the order in forming the corresponding 
translation for the node from the subtrees it 
dominated. In many phrasal matching ap-
proaches, such as constituency-oriented (Kaji et 
al., 1992; Grishman, 1994) and dependency-
oriented (Matsumoto et al., 1993; Watanabe et 
al., 2000; Aramaki et al., 2001), crossing con-
straints are deployed implicitly in finding the 
structural correspondences between pair of repre-
sentation trees of a source sentence and its trans-
lation in target. Here, in our TCT representation, 
we adopted the use of constraint (Wu, 1995) for a 
constituent unit, where the immediate subtrees 
are only allowed to cross in the inverted order. 
Such constraints, during the phase of target lan-
guage generation, can help in determining the 
order in producing the translation for an interme-
diate constituent unit from its subtrees when the 
corresponding translation of the unit is not asso-
ciated in the TCT representation. 
Tree
Source
String {
Onde
1
ficam
2
Adv(1/1/5-6) V(2/2/4)
NP(4/3-6/1-3)
S(2/1-6/1-6)
VP(2/2-6/1-4)
在
4
哪
5
裡
6
更
1
衣
2
室
3
Target
String {
as
3
barracas
4
de
5
praia
6
 
Figure 3. The transfer relationships between the 
sentence-constituents of source language and its 
translation in target language are recorded in 
TCT structure. 
Figure 3 demonstrates the crossing relations 
between the source and target constituents in an 
TCT representation structure. In graphical struc-
ture annotation, a horizontal line is used to repre-
sent the inversion of translation fragments of its 
immediate subtrees. For example, the translation 
substring “更衣室在” of the shaded node, VP, 
can be obtained by inverting the order of the cor-
responding target translations “在” and “更衣室” 
from the dominated nodes V and NP. Therefore, 
such schema can serve as a mean to represent 
translation examples, and find structural corre-
spondences for the purpose of transfer grammar 
learning (Watanabe et al., 2000; Matsumoto et 
al., 1993; Meyers et al., 1998). 
3 Construction of Example Base 
In the construction of bilingual knowledge base 
(example base) in example-based machine trans-
lation system (Sato and Nagao, 1990; Watanabe 
et al., 2000), translation examples are usually 
annotated by mean of a pair analyzed structures, 
where the corresponding relations between the 
source and target sentences are established at the 
structural level through the explicit links. Here, 
to facilitate such examples representation, we use 
the Translation Corresponding Tree as the basic 
annotation structure. The main different and ad-
vantage of our approach is that it uses a single 
language parser to process other than two differ-
ent parsers, one for each language (Tang and Al-
Adhaileh, 2001). 
In our example base, each translation pairs is 
stored in terms of an TCT structure. The con-
struction starts by analyzing the grammatical 
structure of Portuguese sentence with the aid of a 
Portuguese parser, and a shallow analysis to the 
Chinese sentence is carried out by using the Chi-
nese Lexical Analysis System (ICTCLAS) 
(Zhang, 2002) to segment and tag the words with 
a part of speech. The grammatical structure pro-
duced by the parser for Portuguese sentence is 
then used for establishing the correspondences 
between the surface substrings and the inter lev-
els of its structure, which includes the correspon-
dences between nodes and its substrings, as well 
as the correspondences between subtrees and 
substrings in the sentence. Next, in order to iden-
tify and establish the translation correspondences 
for structural constituents of Portuguese sentence, 
it relies on the grammatical information of the 
analyzed structure of Portuguese and a given bi-
lingual dictionary to search the corresponding 
translation substrings from the Chinese sentence. 
Finally, the consequent TCT structure will be 
verified and edited manually to obtain the final 
representation, which is the basic element of the 
knowledge base. 
3.1 The TCT Generation Algorithm 
In the overall construction processes, the task to 
compile the syntactic structure of source sentence 
into the TCT representation by linking the trans-
lation fragments from the target sentence is the 
vital part. The following steps present the com-
plete process to generate an TCT structure for a 
translation example “Actos anteriores à publici-
dade da acção (Publicity of action prior to acts) / 
在訴訟公開前所作之行為”. 
Parsing Portuguese Sentence 
The process begins by parsing the Portuguese 
sentences with a Portuguese parser. The parsing 
result is a phrase structure in terms of bracketed 
annotation. Each bracketed constituent of the 
structure tree is attached with a grammatical 
category. Figure 4 shows the resultant parsed 
structure of the Portuguese sentence. 
(S (N Actos) (AdjP (Adj anteriores) (PP (Prep à)
(NP (N publicidade) (PP (Prep da) (N acção))))))
ParserActos anteriores à publicidade da acção
 
Figure 4. Portuguese sentence is analyzed by a 
linguistic parser, and its output is the phrase 
structure expressed in bracket notation. 
在訴訟公開前所作之行為
在/p 訴訟/v 公開/v 前/f 所/u 作/v 之/u 行為/n
Lexical
Analyser
 
Figure 5. The analyzed lexical items for Chinese 
sentence. 
Analyzing Chinese Sentence 
The construction of TCT structure is fundamen-
tally based on the syntactic structure of Portu-
guese sentence. The finding of translation units 
between the sentences pair is relying on structure 
tree of Portuguese sentence and the sequences of 
lexical words from Chinese sentence. Thus, in-
stead of analyzing the Chinese sentence in deep, 
we analyze the Chinese sentence in the lexical 
level by using the Chinese Lexical Analysis Sys-
tem (ICTCLAS) (Zhang, 2002). Each Chinese 
word is delimited with spaces and assigned with 
a part of speech as illustrated in Figure 5. 
Constructing Correspondence Structure 
for Portuguese Sentence 
After parsing and obtaining the syntactic struc-
ture of Portuguese sentence, next step is to com-
pute the correspondences for the structure against 
the surface strings of the source sentence, which 
includes the corresponding phrase for a constitu-
ent unit in the tree and the corresponding content 
word that headed the constituent unit, both of 
these correspondences are denoted by the se-
quence intervals of the substrings spanning 
across the sentence fragments. In finding the cor-
responding phrasal substrings for subtrees, we 
start associating the lexical words to the corre-
sponding terminal nodes of the structure tree by 
assigning the related offsets to SNODE(n) and 
STREE(n) of the nodes. Then we proceed to next 
upper level constituent units in the tree where the 
corresponding substrings are derived by connect-
ing the lexical words from the nodes in the lower 
level it dominated. Theoretically, if node, N, has 
m daughters, N
1
…N
m
, then the sequence interval 
for N will be STREE(N) = STREE(N
1
) ∪ 
STREE(N
2
) ∪…∪ STREE(N
m
), the interval is 
bounded by spanning nodes of its immediate sub-
trees. To identify the lexical head for a constitu-
ent unit, we use simple rule to determine it by 
considering the grammatical category of the 
phrasal unit, and choose the word that owns the 
same category from the daughter nodes, then as-
sign the interval of chosen to SNODE(N). Figure 
6 shows the structure produced in this stage. 
Actos
1
N(1/1)
anteriores
2
à
3
publicidade
4
da
5
acção
6
S(1/1-6)
AdjP(2/2-6)
PP(3/3-6)
NP(4/4-6)
PP(5/5-6)
Adj(2/2) Prep(3/3) N(4/4) Prep(5/5) N(6/6)
 
Figure 6. The Portuguese correspondence struc-
ture. 
Associating Translation Correspondences 
In this process, we adopt a search for alignments 
between constituent units of Portuguese sentence 
and the corresponding translation fragments from 
Chinese sentence, proceeding bottom-up through 
the tree. It makes use of the information about 
possible lexical correspondences from a bilingual 
dictionary and the grammatical categories of the 
lexical words, tagged in previous stage, to gener-
ate initial candidate alignments. Figure 7 presents 
the initial lexical alignments. 
Actos
N
  anteriores
Adj
  à
Prep
  publicidade
N
  da
Prep
  ac
所
U
 作
V
 之
U
 
ção
N
在
P
 訴訟
V
 公開
V
 前
F
 行為
N
 
Figure 7. Initial candidate alignments of corre-
sponding words. 
Based on the possible word correspondences, 
the associated structure of the Portuguese sen-
tence, together with the grammatical categories 
information, the search proceeds to align phrases 
by gradually increasing length (phrasal corre-
spondences in different levels of constituent tree) 
based on the following criterions. 
First, for any un-aligned words sequence “w
ua
” 
being bounded by aligned words of daughter 
nodes “w
a-left
” and “w
a-right
”, we take the whole 
fragment “w
a-left
w
ua
w
a-right
” (including the bound-
ing words or phrases) as the corresponding sub-
string for the parent node that immediately 
dominates the daughter nodes, such that STC(N) 
= STC(N
left
) ∪ STC(N
right
).  
Second, for the case that the un-aligned frag-
ment is not bounded by any aligned units, our 
approach relies on the assumption that if two set 
of sentence constituents (source and target sen-
tences) are corresponding, their grammatical 
categories as well as the number of constituents 
should be consistent. The essential idea of the 
search is to look for inter levels where the con-
stituent units of the structure of Portuguese sen-
tence and the lexical words in Chinese sentence 
can be projected in one-to-one manner. We use 
the previous example “Onde ficam as barracas 
de praia? (Where are the bathhouses?)/ 
更衣室在哪裡?” to illustrate the searching strat-
egy. Beside the corresponding lexical items, e.g. 
“Onde / 哪裡” and “Ficam / 在”, that can be de-
termined with the aid of a given dictionary, the 
process proceeds bottom-up and searches through 
the tree by considering only the unmatched items 
that if the assumption hold or not. For example, 
at the leaf level, the different numbers of the 
lexical items (“as
Det
, barracas
N
, de
Prep
, praia
N
” 
and “更衣室
N
”) violates the assumption. The 
process repeats the investigation in next upper 
level in the representation structure of Portuguese 
sentence. As illustrated in Figure 8, the alignment 
can be identified only at the level where the 
number and the part of speech of constituent unit 
of Portuguese (“[as barracas
 
de
 
praia]
NP
”) are 
consistent to that of the lexical item in Chinese 
sentence (“[更衣室]
N
”). Consequently, the corre-
spondences between the associated structure of 
Portuguese sentence and the translation frag-
ments of Chinese sentence can be determined and 
established. For any node in the structure which 
has no translation equivalent is assigned with 
“empty (Ø)” interval to STC(N). 
PP(5/5-6/Ø)
Adv(1/1/5-6) V(2/2/4)
NP(4/3-6/1-3)
S(2/1-6/1-6)
VP(2/2-6/1-4)
NP(4/3-4/Ø)
Det(3/3/Ø) N(4/4/Ø) Prep(5/5/Ø) N(6/6/Ø)
Onde
1
ficam
2
as
3
barracas
4
de
5
praia
6
更
1
衣
2
室
3
[N]在
4
[V]哪
5
裡
6
[Adv]
 
Figure 8. Finding the alignment for unbounded 
words. 
Third, for acquiring the crossing constraint for 
a constituent node in the representation tree, 
which is determined by examining the order of 
the translation correspondences of the spanning 
nodes against the sequence of those appeared in 
Chinese sentence. For any node that representing 
Portuguese phrase whose corresponding transla-
tion is derived from its daughters by inverting the 
corresponding translations is denoted by assign-
ing a Boolean value to INVERT(N) attached to 
the node. In graphical annotation, a horizontal 
line is used as a sign for indicating the inversion. 
As demonstrated in Figure 9, the corresponding 
translations of the daughters of node S are 
crossed between the sentences of Portuguese and 
its translation in Chinese. The corresponding 
translation “在訴訟公開前” of its second daugh-
ter appears prior to that “行為” of the first daugh-
ter node in the target translation of Portuguese 
sentence. Hence the inversion property for the 
constituent node in the syntactic structure of 
source sentence is consequently determined.  
S(1/1-6/1-11)
在
1
訴
2
訟
3
公
4
開
5
前
6             
所
7
作
8
之
9             
行
10
為
11
anteriores
2
 à
3
 publicidade
4
 da
5
 acção
6
AdjP(2/2-6/1-6)
N(1/1/10-11)
Actos
1
 
Figure 9. Determination of crossing dependency 
between the translation correspondences 
Finally, in case the representation of TCT gen-
erated in previous process needs further editing, 
an TCT editor can be used to perform the neces-
sary amendment. Figure 10 presents the final 
TCT structure describing a translation example. 
S(1/1-6/1-11)
AdjP(2/2-6/1-6)
PP(3/3-6/1-5)
NP(4/4-6/2-5)
PP(5/5-6/2-3)
N(1/1/10-11) Adj(2/2/6) Prep(3/3/1) N(4/4/4-5) Prep(5/5/Ø) N(6/6/2-3)
Actos
1
anteriores
2
à
3
publicidade
4
da
5
acção
6
在
1
  訴
2
訟
3
  公
4
開
5
  前
6
  所
7
  作
8
  之
9
  行
10
為
11
 
Figure 10. An TCT structure constructed for the 
translation example “Actos anteriores à publici-
dade da acção (Publicity of action prior to acts) / 
在訴訟公開前所作之行為”. 
3.2 Translation Equivalents 
Through the notation of translation correspond-
ing structure for representing translation exam-
ples in the bilingual knowledge base, the 
translation units between the Portuguese sentence 
and its target translation in Chinese are explicitly 
expressed by the sequence intervals STREE(n) 
and STC(n) encoded in the intermediate nodes of 
an TCT structure, that may represent the phrasal 
and lexical correspondences. For instance, from 
the translation example being annotated under the 
TCT representation schema as shown in Figure 
10, the Chinese translation “訴訟” of Portuguese 
word “acção” is denoted by [STREE(n)=6/ 
STC(n)=2-3] in the terminal node. For phrasal 
translation, we may visit the higher level con-
stituents in the representing structure of TCT and 
apply the similar coding information to retrieve 
the corresponding translation for the unit that 
representing a phrasal constituent in a sentence. 
Each TCT structure is being indexed by its nodes 
in the bilingual knowledge base, in order that the 
representation examples can be effectively con-
sulted. 
4 Conclusion 
In this paper, a novel annotation schema for 
translation examples, called Translation Corre-
sponding Tree (TCT) structure, is proposed and 
has been applied to the construction of bilingual 
knowledge base (example base) to be used for the 
Portuguese to Chinese machine translation sys-
tem. The TCT representation provides a flexible 
nature to describe the corresponding relations 
between the inter levels of the structure against 
its substrings in a sentence, in particular the cor-
responding translation fragments (substrings) 
from the target translation sentence are explicitly 
expressed in the structure. We have proposed a 
strategy to semi-automate the example base con-
struction process. A preliminary TCT structure 
for a translation example is first produced by the 
system, then the representation structure can be 
further modified manually through an TCT editor 
to get the final structure.  
Acknowledgement 
The research work reported in this paper was 
supported by the Research Committee of Univer-
sity of Macao under grant CATIVO:3678. 

References 

Mosleh Hmoud Al-Adhaileh, Enya Kong Tang, and 
Yusoff Zaharin. 2002. A Synchronization Structure 
of SSTC and Its Applications in Machine Transla-
tion. The COLING 2002 Post-Conference Work-
shop on Machine Translation in Asia, Taipei, 
Taiwan. 

Eiji Aramaki, Sadao Kurohashi, Satoshi Sato, and 
Hideo Watanabe. 2001. Finding Translation Corre-
spondences from Parallel Parsed Corpus for Ex-
ample-based Translation. In Proceedings of MT 
Summit VIII, pp.27-32. 

Christian Boitet, and Yusoff Zaharin. 1988. Represen-
tation trees and string-tree correspondences. In 
Proceeding of COLING-88, Budapest, pp.59-64. 

Ralph Grishman. 1994. Iterative Alignment of Syntac-
tic Structures for a Bilingual Corpus. In Proceed-
ings of Second Annual Workshop on Very Large 
Corpora (WVLC2), Kyoto, Japan, pp.57-68. 

Hiroyuki Kaji, Yuuko Kida, and Yasutsugu Mori-
moto. 1992. Learning Translation Templates from 
Bilingual Text. In Proceeding of COLING-92, 
Nantes, pp.672-678. 

Yuji Matsumoto, Hiroyuki Isimoto, and Takehito 
Utsuro. 1993. Structural Matching of Parallel 
Texts. 31st Annual Meeting of the Association for 
Computational Linguistics, Columbus, Ohio, 
pp.23-30. 

Adam Meyers, Roman Yangarber, and Brown Ralf. 
1998. Deriving Transfer Rules from Dominance-
Preserving Alignments. In Proceedings of Coling-
ACL (1998), pp.843-847. 

Satoshi Sato, and Magnus Nagao. 1990. Toward 
Memory-Based Translation. In Proceeding of Col-
ing (1990): pp.247-252. 

Enya Kong Tang, and Mosleh Hmoud Al-Adhaileh. 
2001. Converting a Bilingual Dictionary into a Bi-
lingual Knowledge Bank based on the Synchronous 
SSTC. In Proceedings of Machine Translation 
Summit VIII, Spain, pp.351-356. 

Suo Ying Wang, and Yan Bin Lu. 1999. Gramática da 
Língua Portuguesa. Shanghai Foreign Language 
Education Press. 

Fai Wong, Yu Hang Mao, Qing Fu Dong, and Yi 
Hong Qi. 2001. Automatic Translation: Overcome 
the Barriers between European and Chinese Lan-
guages. In Proceedings (CD Version) of First In-
ternational UNL Open Conference 2001, SuZhou 
China. 

Dekai Wu. 1995. Grammarless extraction of phrasal 
translation examples from parallel texts. In Pro-
ceedings of TMI-95, Sixth International Confer-
ence on Theoretical and Methodological Issues in 
Machine Translation, v2, Leuven Belgium, pp.354-372. 

Hua Ping Zhang. 2002. ICTCLAS. Institute of Com-
puting Technology,Chinese Academy of Sciences: 
http://www.ict.ac.cn/freeware/003_ictclas.asp. 
