<?xml version="1.0" standalone="yes"?>
<Paper uid="C94-1003">
  <Title>A Method for Distinguishing Exceptional and G(m(.~L1 .... Examples-' ~ in Example-based Tr~msfer Systems</Title>
  <Section position="3" start_page="39" end_page="40" type="metho">
    <SectionTitle>
2 Mechanism for dealing with
</SectionTitle>
    <Paragraph position="0"> i, rans\[;tLion pattern SimTlNn calculates the similarity between a subgraph of an input structure and the source part of a translation pattern on the basis of both the structural similarity and the similarity of the lexical-forms of corresponding nodes. For instance, the distance (the inverse of similarity) between two Japanese lexical-forms is expressed by the difference of their values in a Japanese thesaurus called Bunrui-Goi-lIyou \[5\] 3 as follows:</Paragraph>
    <Paragraph position="2"> where bghcode(w ) is the code vMue in the Bunrui-Goi-Hyou, bghma:c is the maximal difference of the bghcodes, and 6 is a penalty value incurred when wl and w2 are not identical. This equation is used for lexical-forms in general translation patterns. If one is a lexicM-form which requires exact-match in an exceptional translation pattern, then the distance is calculated as follows:</Paragraph>
    <Paragraph position="4"> aBunrui-Goi-IIyou is a Japanese thesaurus consisting of large trees for nominals, adjectives, and verbs. Each node is assigned a unique nmnber. Similar concept words are locattxl in similar positions (or assigned similar numbers) in these trees.</Paragraph>
    <Paragraph position="5"> A lexical-forni has a distinctive fea.tnre that makes it possible to determine which equation should be used hi cMculating similarity I if one of two le.xlcal-forms is expressed by a single-quoted string, then the distance between the lexical-forms is calculated by using the second equation; on the other hand, if both lexical-forms are expressed by double-quoted strings, then their distance is calc:nlated by using the first equation.</Paragraph>
    <Paragraph position="6"> Thus, an exceptional translation pattern is distinguished by having nodes whose lexicM-forrns are single-quoted strings in its source part, while a general translation pattern is distinguished by having nodes whose lexicM-fi~rms are all double-quoted strings in its source part. Not MI nodes in the source part of an exceptional translation pattern are necessarily single-quoted strings; single-quoted string nodes and don bh+-quoted string nodes may be mixecl in a translation pattern, ht Figure 2, (tpl) is an exceptional tr;ulslation pattern and (tp2) is a general translation pattern. Note tt~tt the root node of the Japanese part is the only single-quoted string in (tpl), and it matches only an input whose root node is 'kyouyoLIsHru. ~ By using this distinction of lexical-forrns, we e~n integrate exceptionality handling into the similarity calculation framework without separating this task as a pre process or post-process. '  &amp;quot;kyouyou- &amp;quot; ..... &amp;quot;share&amp;quot; suru&amp;quot; i dob\] l wo &amp;quot;use&amp;quot; +the ~ postmod &amp;quot;kuruma&amp;quot; * ....... (&amp;quot;ear&amp;quot;) &amp;quot;car&amp;quot; +of (tpl) &amp;quot;tsukau&amp;quot; &amp;quot; ............ use&amp;quot; 1degdegdeg, &amp;quot;kuruma&amp;quot; .............. car&amp;quot;</Paragraph>
  </Section>
  <Section position="4" start_page="40" end_page="40" type="metho">
    <SectionTitle>
3 Method for identifying ex-
</SectionTitle>
    <Paragraph position="0"> ceptional translation patterns null For iriost peol)le , an exceptional translation pattern is likely to recall a pattern of translation for an i/liomatic or colloquial expression, hi generM, ;in idiomatic translation pattern is a translation pattern whose target part is markedly different from that of translation patterns whose sC/)urce parts am similar to that of the idiomatic pattern. Froni the viewpoint of the transfer process, what we would like to identi\[y are translation patterns that may have side-effects when they are selected instead of general translation patterns. We call such translation patterns exceptional travsIation pattern.s. According to this defi nltlon, exceptional translation patterns are not restricted to idiomatic patterns, in fact, more translatlon l)atterns other than idiomatic ones fall into this category. Here: we classify exceptional translation patterns into the following two categories: t Extra-Exceptional Translation l'atterns: These have some. extra elements hi the. target part in addition to those in similar traimhttimi patterns.</Paragraph>
    <Paragraph position="1"> i Intra-Exceptional &amp;quot;\]'rans\[atlon \]&gt;atterns: These are almost same ms similar translation patterns, but several target words are different.</Paragraph>
    <Paragraph position="2"> of 0xeeptimial triulslatlon patterns When exceptional translation patterns are \[olind~ it is hnportant to know whether two translatiml patterns are e(lUivMent or not. '\]'herefore&gt; equivalent translation plctterns are defined as follows: (liven two dependency structures dl and d2, then they lore called equivalent if and only if tiiey are strllerurally identicM and correspmiding nodes have the similar seinantic code. 4 }&amp;quot;urther~ given two trailslath,n patterns tp, = (si,ti,m,) tp2 = (s2,t2,m2), where .~i is ~L so)lr('e l)art, ti in a target part, an(I mi i~ a mapping from .~'i to iT, then these two translation patterns ;ere called equivalent if they satisfy the following conditions:  (1) Both sou roe parts axe equivalent&gt; and both targ~t parts are strilctllrally identical.</Paragraph>
    <Paragraph position="3"> (~) 'l'he roots of l 1 a.nd 17 are the sallle strhlg.</Paragraph>
    <Paragraph position="4"> (3) For each ,m(le n hi .+~, ',n~(n)is o,,e of transhttion words of n.</Paragraph>
    <Paragraph position="5"> (4) t,'o~ each ,~o,le ,, in ,&amp;quot;2, ',,.,('n) is one or translation  words of n.</Paragraph>
    <Paragraph position="6"> The. algorithm for identi\[yhlg exceptlonal trluislation patterns is as follows: ,I \]?Of ili.~ltltlll It~ tiil~ :'ll!lllitllt, iC code ill JIL|llllll!~\[~ \[~+ ~llllrll\[-(loi-Hyou code. The extent to whh:h two words are determhw, d to I,c similar is *also a p~ranleter. It may vary according to the system. In this liltper, two words iu't~ deternllncd to be similar if they have the ~anle senuu~tic c,Me.</Paragraph>
    <Paragraph position="7"> 4&amp;quot;/ Step 1 Divide translation patterns into sew~ral groups, each of which consists of equlwdent translation patterns.</Paragraph>
    <Paragraph position="8"> Step 2 For each pair of distinct translation pattern groups gl and g~, if any pattern of 9t is equivalent to any pattern of g2 other than nodes governed by the root of the source l)art, tlmn the translation patterns in gl arid 92 are marked gener'~L null Step 3 ~br each pair of distinct translation pattern groups gx and g2, if&amp;quot; the source part of any pattern (pl) of gl is equivalent to the source part of any pattern of g2, but target parts of them are not struetnrally identical, because Pl ha.s extra elements~ then the translation patterns of gl are marked extm-exeeptionaL Step 4 For each non-exceptional translation pattern group gl, if there is another general translation pattern group g~ such that any pattern (Pl) of gl is equivMent to any pattern of g2 other than the root node in the target part of Pt, then the translation patterns of gt are marked itth'aexceptional. null Step 2 identifies possible general translation patterns if they are used in a relatively wide range of'words, because in general an exceptional pattern is restricted in the usage of words. This approach, however, is not perfect rot identif,ying general translation patterns, becanse there in ~t c~use such that the exccptionality derives from a single special word. Therefore, in the next step, checking does riot exclude these possible general translation patterns. Step 3 identities extra-exceptional translation patterns by checking the structure of the target part. Step 4 then identifies intra-exceptional ones by comparing the mot node in the target part with the root nodes in the target part of possible general translation patterns. The reason why this comparison is restricted to possible general translation patterns is that intra-excepti(n,d translation patterns have si(h~efrects only when they are similar to general translation patterns.</Paragraph>
    <Paragraph position="9"> Figure 3 shows an example of the identiflcation of exceptional translation patterns, in which the Japanese verbs &amp;quot;kyouyousuru&amp;quot; and &amp;quot;tsukau&amp;quot; haw.' the same bghcode, and the Japanese nouns &amp;quot;kuruma,&amp;quot; &amp;quot;denwa~&amp;quot; and &amp;quot;mahou&amp;quot; have different bghcodes, on the other hand, &amp;quot;kuruma&amp;quot; and &amp;quot;jitensyd' have the same bg|,eode. First, step 1 divides tImse translation patterns into four groups: group 1 c.onsists of (tpl), group 2 consists of (tp2) and (tp3), group 3 consists of (tp4), and group 4 consists of (tp5). Step 2 identifies group 2 and 3 as general translation patterns, because &amp;quot;kuruma&amp;quot; and &amp;quot;denwa&amp;quot; have different bghcodes. Subsequently, step 3 identifies (tpl) as an extra-exceptional translation pattern, beci~use (tpl) has extra elements &amp;quot;the use of&amp;quot; for (tp~). Further, step 4 identifies (tpS) as at, iutra-exceptional translation pattern, because (tp5) is equivalent to the general translation patterns (tp2), (tp3) and (tp4), other than &amp;quot;use&amp;quot; and &amp;quot;practice&amp;quot; in the root nodes of the target parts.</Paragraph>
  </Section>
  <Section position="5" start_page="40" end_page="40" type="metho">
    <SectionTitle>
4 Experiments
</SectionTitle>
    <Paragraph position="0"> We have tested the almve-nientioned algorithm with translation patterns in a Japanese-to-English transfer dictionary that was previously used in our laboratory. For each bghcode, we. collected translation patterns such that the root of the source part has the. code. and a.pplied the algorithm to tim translation pattern set of each category. Table 1 shows the resulting top 10 categories with respect to tt,e total number of occurrences. In most categorles, more than 90% of translation patterns were identified as exceptional. The reason for the lopsidedness of, this result is that tl,e translation patterns described in the pr(~ vious transfer dictionary were almost all exceptional eases that conhl not be. de.all with by the default procedures coded in the transfer module. Therefore, this result indicates that the ~dgorithm is able to idenitfy exceptional translation patterns correctly.</Paragraph>
  </Section>
class="xml-element"></Paper>