A Cross-Lingual Grammar Model
and its Application to JapaneseSpanish Machine Translation
Manuel Medina González and
Hirosato Nomura
Kyushu Institute of Technology
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
Introduction

Current Machine
Translation Systems
output incorrect
sentences when
translating from
Japanese to Spanish.
http://www.worldlingo.com/en/products_services/worldlingo_translator.html
Introduction

The reason is
because English is
used as intermediate
language, thus,
leading to loss of
grammatical
information due to
the differences
between the
languages.
http://www.worldlingo.com/en/products_services/worldlingo_translator.html
Introduction

The idea is simple: To translate directly
from Japanese to Spanish.
子供は公園で遊ぶ

El niño juega en el
parque
In order to accomplish this, the way the
analysis is performed must be adapted
to support Spanish features
Introduction

We base our model on ALT J/E Machine
Translation System Model, with some
modifications.
Model
Source
Language
Direct Method
Corpus-based
Translation
Target
Language
• Noun features
• Determiner
• Subjunctive
Mood
Analysis
Generation
Transfer Method
Conversion
Intermediate
Language
(PIVOT)
Model
Determining the missing
information by predicting the
result as we analyze the
sentence.
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
Spanish Features: Nouns

Gender
Table: Feminine.
Book: Masculine.

Number
Table: Singular
Tables: Plural
Noun’s features decide
almost all the possible
changes a Spanish
sentence can suffer.
Spanish Features: Nouns
あのテーブルは汚い。拭いておきましょう。
女性
あの → 女性形
汚い → 女性形
Zero代名詞 → 女性形
Semantic Categorization

ALT J/E Semantic Categorization (2710
different, non-exclusive categories).
Spanish Features: Adjectives



As in English, only one category exists.
2 verbs mainly used: “Ser” and “Estar”.
Both are equivalent to English “To be”
verb.
The meaning is different depending on
the verb used.
Yo soy feliz
私は幸せだ
Yo estoy feliz
Spanish Features: Adjectives

Creation of categories




Temporary state: Sad
Permanent feature: Boring, interesting
Weather
Weather category is necessary because
other 2 verbs are used: “Tener” and
“Hacer”.
What you feel
Tengo calor
暑い
Hace calor
The weather
Spanish Features: Adverbs

Thinking of them as they are
in Spanish, we create
categories as in this
language:










Place
Time
Mode
Quantity
Order
Affirmation
Denial
Doubt
Addition
Exclusion
Spanish Features: Verbs

Tenses



Japanese: 3 (Present, past, future)
Spanish: 16
Conjugations


Different conjugation for each person in
each tense.
A conjugator system can be made for
regular verbs, but there are too many rules
to consider.
直説法・現在
juego
juegas
juega
jugamos
juegan
juegan
直説法・完了過去
jugué
jugaste
jugó
jugamos
jugaron
jugaron
直説法・不完了過去
jugaba
jugabas
jugaba
jugábamos
jugaban
jugaban
直説法・未来
jugaré
jugarás
jugará
jugaremos
jugarán
jugarán
直説法・可能形
jugaría
jugarías
jugaría
jugaríamos
jugarían
jugarían
命令法
juega
jueguen
接続法・現在
juegue
juegues
juegue
juguemos
jueguen
jueguen
接続法・不完了過去
jugara
jugaras
jugara
jugáramos
jugaran
jugaran
現在分詞
jugando
過去分詞
jugado
直説法・現在完了
he jugado
has jugado
ha jugado
hemos jugado
han jugado
han jugado
直説法・直前過去
hube jugado
hubiste jugado
hubo jugado
hubimos jugado
hubieron jugado
hubieron jugado
直説法・未来完了
habré jugado
habrás jugado
habrá jugado
habremos jugado
habrán jugado
habrán jugado
直説法・大過去
había jugado
habías jugado
había jugado
habíamos jugado
habían jugado
habían jugado
直説法・過去未来
形
habría jugado
habrías jugado
habría jugado
habríamos jugado
habrían jugado
habrían jugado
接続法・完了過去
haya jugado
hayas jugado
haya jugado
hayamos jugado
hayan jugado
接続法・大過去
hubiera jugado
hubieras jugado
hubiera jugado
hubiéramos
jugado
接続法・未来形
hubiere jugado
hubieres jugado
hubiere jugado
hubiéremos jugado
hubieren jugado
日本語:遊ぶ(五段活用)
遊ばない、遊びます、遊ぶとき、遊べる、
遊ぼう。
スペイン語: Jugar
種類:語幹が変わる規則動詞。 原形:
Jugar
Reflexive Verbs


A verb is reflexive if the action returns
to its performer.
There are verbs in Japanese that can be
reflexive and non-reflexive at the same
Non-Reflexive
time.
Reflexive
私は車を洗う
私は顔を洗う
顔
Reflexive Verbs

Creation of “Has-a” relationships to
determine whether if a verb must be
treated as reflexive.
目
人間
顔
鼻
Translation Rules

Based on ALT J/E Translation Rules.





Verb
Particles used in special cases
Categories of the expected nouns
accompaining the particles
Translation of the verb in each case.
Indication if the verb must be treated as
reflexive.
「乗る」
<動物>に乗る = Montar en <動物>
<交通機関 | 乗り物>に乗る = Subir(R)
a <交通機関 | 乗り物>
...
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
Voices

Passive: れる、られる




Normal passive
Indirect object reference
Passive Reflexive
Causative: せる、させる


Coercive
Permissive
Voices: Model
Japanese Sentence
Identify Voice
Voice,
Structure
Passive,
Causative
Predict Result
Add or change elements
Analyze elements
…
Pronouns
Mood...
Not explained
deeply here due to
the time limitation
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
System

JUMAN
Named “JEMS”: Japanese Español
Machine translation System.
KNP
Dictionary
JEMS Core
Translation
Rules
=
Translated
Sentence
Semantic
Categories
JEMS
Tests and Results



JEMS compared against Worldlingo.
Sentences taken from books like
“Momotaro”, “Megane usagi”, “3 nen
netaro” etc.
Human-Translating the sentences, then
inputting them into the systems and
checking the output.
Tests and Results
Evaluation
JEMS
WL
Correct
58%
24%
Structure Errors
(acceptable)
18%
26%
Conj. Errors
(acceptable)
12%
11%
Incorrect
(non-acceptable)
6%
25%
Other Errors
6%
0%
Tests and Results
Possible Expected
Outputs
Input
いよいよ春がやってきた
Finalmente la primavera llegó.
Finalmente la primavera ha llegado.
Obtained Output
Resorte usted cada vez más
Errors:
• Lack of verb
• Incorrect subject
• Incorrect structure
春 → Spring
Spring =
1. Primavera
2. Resorte
Analysis is not complete
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
Summary




Indirect Translation from Japanese to Spanish
is not enough.
Model based on thinking of the translated
sentence since the analysis starts.
Presented just a small part of the analysis
necessary to translate into Spanish
Developed a prototype system “JEMS” to test
the model. Compared against an existent
translation system.
Outline


Introduction
Spanish Features and Considerations
when translating.





Parts of speech
Voices
System
Summary
Conclusions
Conclusions



Japanese-Spanish Machine Translation is just
beginning. There are still many issues to be
solved.
Need to make the model bigger in order to
analyze longer sentences.
Once this model is finished, it can become
the basis for other research about Machine
Translation between Japanese and Romance
Languages.
Conclusions
Spanish
Me llamo Taro
Me chiamo Taro
私は太郎です
Italian
Portuguese
Me chamo Taro
French
Je m'appelle Taro
Voices: Passive

Normal Passive

Very much like English Passive voice:
Subject + Verb + Object
Subject + “Ser” Verb + Verb’s Past Participle + “por” preposition + Agent
子供はボールを蹴った
ボールは子供に蹴られた
Voices: Passive
ボールは子供に蹴られた
1. Identify the agent (子供に).
2. Identify the subject and its
features. (ボールは→女性) .
Past participle must match
these features
3. Use translation rules to get
the appropiate verb
translation.
4. Change the past participle
according to subject features.
Translated
Sentence
Voices: Passive

Indirect object reference
私は財布を盗まれた
The subject of the Spanish sentence is neither “I” nor
“Wallet”. “I” is the indirect object in the Spanish
translated sentence.
Indirect
Object
私は財布を盗まれた
Weird Japanese
(ZERO)は私に財布
を盗んだ。
Voices: Passive
私は財布を盗まれた
1. No agent in the sentence. Rewrite it to “Weird Japanese”
form (ZERO)は...
2. Get correct verb translation from translation rules.
3. Use conjugation for “they”
Past participle
is not used
Translated Sentence
Voices: Passive

Passive Reflexive.
Identified only in some patterns:


~では・には・・・される
Sentences that translated into Spanish
don’t have subject, the agent is present in
the sentence and the verb is 「考える」、「思
う」、「言う」
日本では日本語が話される
Voices: Passive
日本では日本語が話される
1.
2.
Use Reflexive Pronoun “se”.
Use “Singular 3rd. Person”
conjugation
Translated Sentence
Voices: Causative



2 cases: Coercitive Sentences and
Permissive Sentences.
Sentences are translated differently
depending on if the verb is intransitive
or not.
Possible use of subjunctive mood in the
translated sentence.
Descargar

A Cross-Lingual Grammar Model and its Application to