Anna Zaretskaya - ESR 1 UMA

12
what do translators want? Investigating translators’ requirements for translation technologies Anna Zaretskaya May 6, 2016 Rome, Italy

Transcript of Anna Zaretskaya - ESR 1 UMA

Page 1: Anna Zaretskaya - ESR 1 UMA

what do translators want?Investigating translators’ requirements for translationtechnologies

Anna ZaretskayaMay 6, 2016

Rome, Italy

Page 2: Anna Zaretskaya - ESR 1 UMA

Goals

Usersurvey

Translators’needs

Exis,ngtools

Evalua,on Improvements

1

Page 3: Anna Zaretskaya - ESR 1 UMA

Survey content

CAT tools - usage- best and worst features

Machine Translation - usage- integration in CAT- languages and domains

Corpora - using- compiling- special tools

Terminology - management- extraction

2

Page 4: Anna Zaretskaya - ESR 1 UMA

What is new?

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

TM MT TerminologyManagement

Corpora Corporaprocessing

MeLLANGE2006 TM2006 SDL2009 TTC2010 TranslaEonTechnologies2010 AZ2014

3

Page 5: Anna Zaretskaya - ESR 1 UMA

Features of CAT tools

-Sentencealignment-Automa0chandlingoftags

-Real-0mepreview-EditTM-Autopropaga0on

-OCR-Speechrecogni0on

“Justmakeitsimple”

4

Page 6: Anna Zaretskaya - ESR 1 UMA

Machine Translation

∙ 36% of MT users∙ 74% find good MT useful

∙ 1/3 of respondents did notknow if they had MT in theirCAT tool

∙ one of the least usefulfeatures in CAT

5

Page 7: Anna Zaretskaya - ESR 1 UMA

Post-editing of MT

Can PE be a way to improve the MT experience for a professionaltranslator?

MT errors in post-editing

1. take into account when evaluating MT output2. automatically post-edit certain errors3. highlight most difficult errors

6

Page 8: Anna Zaretskaya - ESR 1 UMA

Corpora

- 15% of corpora users

- compiling corpora is time consuming (57%); not familiar with anyspecial tools (34%)

but:

concordance search is favourite feature

∙ make it easy to incorporate more resources in concordance search(not only TM)

∙ parallel and comparable corpora

7

Page 9: Anna Zaretskaya - ESR 1 UMA

Terminology

Appears both in the favourite and most-hated features =>

terminology management is necessary, but the current tools are notgood enough.

- Study different steps in the terminology management process tomake it easier

8

Page 10: Anna Zaretskaya - ESR 1 UMA

Bibliography

Costa, H., Corpas-Pastor, G., Seghiri, M., and Zaretskaya, A. (2016). Nine terminology extraction tools - Are they useful for translators?Multilingual.

Zaretskaya, A. (2016). A quantitative method for evaluation of cat tools based on user preferences. In Proceedings of the AELFE XVInternational Conference. University of Alcalá.

Zaretskaya, A., Corpas Pastor, G., and Seghiri, M. (2015a). Translators’ requirements for translation technologies: a user survey. InCorpas-Pastor, G., Seghiri-Domínguez, M., Gutiérrez-Florido, R., and na, M. U.-M., editors, Nuevos horizontes en los Estudios deTraducción e Interpretación (Trabajos completos) / New Horizons in Translation and Interpreting Studies (Full papers) / Novoshorizontes dos Estudos da Tradução e Interpretação (Comunicações completas), pages 247–254, Malaga, Spain. AIETI, Tradulex.

Zaretskaya, A., Corpas-Pastor, G., and Seghiri, M. (2016a). Corpora in computer-assisted translation: a users’ view. In Corpas-Pastor, G. andSeghiri, M., editors, Corpus-based Approaches to Translation and Interpreting: from theory to applications. Peter Lang, Frankfurt.

Zaretskaya, A., Corpas-Pastor, G., and Seghiri-Domínguez, M. (2016b). A quality evaluation template for machine translation. TranslationJournal.

Zaretskaya, A., Pastor, G. C., and Seghiri, M. (2015b). Integration of machine translation in CAT tools: State of the art, evaluation and userattitudes. SKASE Journal for Translation and Interpretation, 8(1):76–88.

9

Page 11: Anna Zaretskaya - ESR 1 UMA

Thank you! Questions?

10

Page 12: Anna Zaretskaya - ESR 1 UMA

From the horse’s mouth

Agencies using MT are … only interested in money, not in quality and it is our duty to inform end clients about this greedy attitude.

[CAT tools] should be made simple to operate for a linguist not for an IT specialist

I would like to see more compatibility between systems

Unfortunately, many people think that Google translator, or Babelfish, are viable alternatives to paying a human translator.

Make everything easier and easier by making the usage limitless. Thanks.

I feel there is industry pressure now for more translation work to be done on CAT software so perhaps I am a bit of a dinosaur in myattitude!

CAT or TMs are incomprehensible and un-learnable

TM tools are just stealing knowledge from translators and reducing their earnings. I hate them. Thank you.

This survey teaches me that all those tools exist! I wake up to a new world of possibilities!

It would be useful t have a tool that adapts to my ”style”... not only in its terminology but also in its grammatical twists

11