i evaluation of free online machine translations for croatian-english and english-croatian language...
TRANSCRIPT
I
Evaluation of Free Online Machine Translations
for Croatian-English and English-Croatian Language Pairs
Sanja Seljan, [email protected] of Zagreb - Faculty of Humanities and Social Sciences,
Department of Information Sciences, Croatia
Marija Brkić, [email protected] of Rijeka, Department of Informatics, Croatia
Vlasta Kučiš, [email protected] of Maribor, Department of Translation Studies, Slovenia
FF Zagreb – Informacijske znanosti
I
Aim
Text evaluation from four domains (city description, law, football, monitors)
Cro-Eng - by four free online translation services (Google Translate, Stars21, InterTran and Translation Guide)
En- Croatian - by Google Translate Measuring of inter-rater agreement (Fleiss kappa) influence of error types on the criteria of fluency and
adequacy Pearson’s correlation
I
I. Introduction
II. MT evaluation
III. Experimental study Translation tools Test set description Evaluation Error analysis Correlations
IV. Conclusion
I
I INTRODUCTION
increased use of online in recent years, even among less widely spoken languages
Desirable: moderate to good quality translations
evaluation from the user's perspective
Tools and evaluation mainly for widely spoken languages
Possible use: gisting translations, information retrieval, i.e. question-answering systems
1976 Systran - first MT for the Commission of the European Communities + online tool + different versions
1997 - first online translation tool - Babel Fish using Systran technology
Important: realistic expectations
I
Studies for popular languages
Considerable difference in the quality of translation dependent on the language pair
2010 - German-French (GT, ProMT, WorldLingo)
2011- three popular online tools
2006 - Spanish-English (introductory textbook)
2008 – 13 languages into English (6 tools: BabelFish, Google Translate, ProMT, SDL free translator, Systran, World Lingo)
I
MT evaluation – important in research and product design measure system performance identify weak points and adjust parameter settings language independent algorithms (BLEU, NIST) Better metric – closer to human evaluation
need for qualitative evaluation of different linguistic phenomena
I
II EXPERIMENTAL STUDY
evaluation of free online translation services (FTS) – from user’s perspective
undergraduate and graduate students of languages, linguistics and information sciences attending courses on language technologies at the University of Zagreb, Faculty of Humanities and Social Science
Test set description texts 4 domains (city description, law, football, monitors) Cca 7-9 sentence per domain (17.8 word/ sent.) Cro-En, En-Cro
I
EvaluatorsCro-En: 48 students, final year of undergraduate and graduate levelsEn-Cro: 50 students, native speakers75% of students attended language technology course(s)
3 3.57
012345
Croatian in general
Average grades for free language resources on the Internet
Evaluation – before pilot study
I
Croatian tools/resources Tools/ resources in general
3.683.9
3.022.7
0
1
2
3
4
5
Average
Systran Google Translate
InterTran Translation Guide
3.142.49 2.45
3.54
0
1
2
3
4
5
Average
Online dictionariesTerminology databases Translation memories Google Translate
I
Desirable tools/ resources of appropriate quality
0.00% 20.00% 40.00% 60.00% 80.00% 100.00%
online dictionaries
glossaries
term bases
translation memories
MT systems
speech-to-text system
I
Evaluation
Manual evaluation fluency (indicating how much the translation is fluent in the target language) adequacy (indicating how much of the information is adequately transmitted)
evaluation enriched by translation errors analysis
−morphological errors,
−untranslated words
−lexical errors and word omissions
−syntactic errors
I
Tools
Cro-En translations Google Translate (GT) - http://translate.google.com Stars21 (S21) - http://stars21.com/translator InterTran (IT) -
http://transdict.com/translators/intertran.html Translation Guide (TG) - http://www.translation-
guide.com
En-Cro translations obtained from Google Translate
I
Google Translate translation service provided by Google Inc.statistical MT based on huge amount of corporaIt supports 57 languages, Croatian since 2008
S21 service powered by GTtranslations not always the same
InterTranpowered by NeuroTran and WordTransentence-by-sentence and word-by-word
Translation Guidepowered by ITDifferent translations
I
Results - Cro-En
either low grades (TG and IT) or high grades (S21 and GT), in comparison to the average value (3.04)
S21(4.66) : GT (4.62) – city description, legal GT – football, monitors Best average result – legal domain, then monitors and football Lowest – city description (the most free in style)
1
2
3
4
5
City Law Football Monitors
Stars21 Google Translate
InterTran Translation Guide
I
Results - Cro-En
- En-Cro - lower average results than the reverse direction: football (3.75 : 4.84), law, monitors
- Higher average grade in city description (shorter sentences, mostly nominative constructions, frequent terms)
- Football domain - specific terms, non-nominative constructions
I
Error analysis
En-CroTranslations offered by GT and S21 are very similar, although not identicalTG and IT – difference in number of untranslated wordsTG does not recognize words with diacritics
Cro-Enthe highest number of lexical errors, including also errors in style (av. 2.44 ) Untranslated words (1.83), morphological (1.75), syntactic errors (1.38)Lowest score, highest number of errors - football domain (mostly lexical errors and untranslated words)best score – in city description domain (lexcial errors)Lowest no. errors – legal domain (evenly distributed)
I
Morphological errors – mostly in domain of monitors, the smallest no. in city desription (dominant value 1)
Untranslated words - by far mostly in the football
translation grades - mostly influenced by untranslated words
Dominant values Morphological errors: 1 in city description and monitors, 3
in the legal and football Lexical errors: 1 in city description , others higher untranslated words - 1 in all domains syntactic errors - 1 in all domains but football (2-3)
I
Pearson’s correlation
smaller number of errors augments the average gradecorrelation between errors types and the criteria of fluency and adequacy
fluency - more affected by the increase of lexical and syntactic errors,
adequacy is more affected by untranslated words
I
Fleiss' kappa
for assessing the reliability of agreement among raters when giving ratings to the sentences
Indicating extent to which the observed amount of agreement among raters exceeds what would be expected if all the raters made their ratings completely randomly.
Score - between 0 and 1 (perfect agreement)
0.0-0.20 slight agreement N – total of subjects 0.21-0.40 fair agreement n – no. of raters per subject 0.41-0.60 moderate agreement i – extent to which raters
0.61-0.80 substantial agreement agree on i-subject 0.81-1.00 almost perfect agreement j - categories
N
i
k
jij Nnn
nNnP
1 1
2 )()1(
1
I
relatively high level of the agreement among raters per domain and per system in Cro-En translations moderate 0.41-0.60 (for IT translation service), substantial agreement 0.61-0.80 (S21 and GT) perfect agreement 0.80-1.00 (TG – the worst tool)
En-Cro translations - inter-rater agreement per domain lowest level of agreement has been detected in the
domains of football and law (from 0.4-0.49 fair & moderate) – larger and more complex sentences
substantial agreement (0.61-0.80) – in city description level of inter-rater agreement is lower for En-Cro
translations in all domains
I
Conclusion
evaluation study of MT in 4 domains Cro-En – 4 free online translation services En-Cro translations – by Google Translate
Evaluator’s profile high interest in use of translation resources and tools Critical evaluation
System evaluation perfect agreement in the ranking of TG as the worst translation
service substantial agreement is achieved for S21 and GT services moderate agreement is shown for IT, which has performed
slightly better than TG.
I
Cro-En translations S21 and GT ( 4.63 to 4.84) - football, law and monitors city description - Cro-En lower than in En-Cro
En-Cro direction – by GTlower grades than in the opposite direction (specific terms, non-nominative constructions, multi-word units)Except city description domain - containing mostly nominative constructions, frequent words, no specific terms
Error analysis translation grades are mostly influenced by untranslated words (especially the criteria of adequacy)morphological and syntactic errors reflect grades in smaller proportion (fluency) ,
I
Google Translate serviceused in both translation directionsharvesting data from the Web, seems to be well trained and suitable for the translation of frequent expressionsDoesn’t perform well where language information is needed, e.g. gender agreement, in MW expressions
Further research Better quantitavie analysis per domainmore detailed analysis of specific language phenomena
I
Evaluation of Free Online Machine Translations
for Croatian-English and English-Croatian Language Pairs
Sanja Seljan, [email protected] of Zagreb - Faculty of Humanities and Social Sciences,
Department of Information Sciences, Croatia
Marija Brkić, [email protected] of Rijeka, Department of Informatics, Croatia
Vlasta Kučiš, [email protected] of Maribor, Department of Translation Studies, Slovenia
FF Zagreb – Informacijske znanosti