Translation for Massive Open Online Courses

Livrables

A plan describing the quality and risk management procedures to be established within the project including procedures for the mutual quality control among the partners

Initial integration report

Report on the integration of the first prototype machine translation system into the Iversity platform.

Initial system architecture

A report on system architecture and requirements analysis, that includes the end user interviewing setup and outcome, the specification decisions for the first and second system prototypes, a detailed description of the system architecture design for both prototypes.

Scientific and commercial exploitation report

A report on the scientific and commercial exploitation of the results achieved within the project

Crowdsourcing platform configuration

A report on the configuration process of the crowdsourcing platform The configuration will be implemented for all three crowdsourcing activities by breaking them down into simple comprehensible tasks Precautions to ensure participation control will be described Instruction manuals for all three activities will be included that contain comprehensible examples of good practice to follow and of bad practice to avoid

Report on crowdsourcing activity 3

A report on the setup, the participation statistics, the size of annotated data, the agreement ratios of the third crowdsourcing activity, as well as its evaluation results.

Report on Crowdsourcing Activity 1

A report on the setup, the participation statistics, the size of translated data, the agreement ratios of the first crowdsourcing activity, as well as its evaluation results.

Impact and continuation plans for TraMOOC services and software

Measures and plans in order to ensure that the services and language tools developed within TraMOOC will remain available after the lifetime of the project. Analysis of the current state and strategic planning of the development community around TraMOOC created open source software.

Final system architecture

A report on system architecture and requirements analysis, that includes the end user interviewing setup and outcome, the specification decisions for the first and second system prototypes, a detailed description of the system architecture design for both prototypes.

Report on crowdsourcing activity 2

A report on the setup, the participation statistics, the size of translated data, the agreement ratios of the second crowdsourcing activity, as well as its evaluation results.

Report on the human and automatic second stage translation evaluation

A report that will include detailed qualitative and quantitative evaluation results of the second stage translation output as well as their comparative analysis across translation models, across language models, across languages and across text types.

Communication pack for the project legacy

This deliverable will include the dissemination materials which will be used by all partners in future events performed after the end of the project’s lifetime. The produced dissemination material (e.g. brochure, etc.) will be mainly focused on providing information about the translation platform and its functionalities.

Report on the human and automatic first stage translation evaluation

A report that will include detailed qualitative and quantitative evaluation results of the first stage translation output as well as their comparative analysis across translation models, across language models, across languages and across text types.

Report on the implicit translation evaluation

A report on the topic identification and sentiment analysis tasks and the implicit translation evaluation process and its results.

Final Plan for the Dissemination and Exploitation of the project results

The aim of this deliverable is to report on the performed dissemination and exploitation activities of the project. Based on the initial Dissemination and Exploitation plan, this deliverable will summarize the consortium’s strategy and concrete actions to disseminate, exploit and protect the foreground generated by the project.

Report on machine translation 3

Report on system development (incl. Prototype 3).

Dissemination and Exploitation plan of the project results

The aim of this deliverable is to report on the planned and performed dissemination and exploitation activities during the first year of the project duration. Based on the preliminary Dissemination and Exploitation included in the proposal, this deliverable will summarize the consortium’s strategy and future actions to disseminate, exploit and protect the foreground expected to be generated within the project.

Language resources and tools management plan

A plan for the management of language resources and tools across the different tasks and work packages of the project

Report on the infrastructure

A report on the collected parallel data from the MOOC partner and from the crawling process, presenting cleaning, formatting and standardization processes and detailed text statistics per text genre and per targeted language. Also, the report will include the description of the exploration process for available tools and resources, as well as the solutions applied for their bootstrapping, adaptation and enhancement.

Initial communication pack

The initial communication pack will include the project logo poster leaflet presentation service mockup demo video and deliverable templates

Intermediate Dissemination and Exploitation plan of the project results

The aim of this deliverable is to report on the planned and performed dissemination and exploitation activities during the first two years of the project duration. Based on the preliminary Dissemination and Exploitation delivered in M12 of the project, this deliverable will summarize the consortium’s strategy and foreseen actions to disseminate, exploit and protect the foreground generated or/and expected to be generated within the project.

Report on machine translation 2

Report on system development (incl. Prototype 2)

Final field testing report

Report on integration efforts of the final system and final field testing of the comprehensive system.

Report on machine translation 1

Report on system development (incl. Prototype 1)

Report on available MT structure

An overview of the MT resources available for each target language and analysis of the issues those pose for the MT system developed within this project.

Initial field testing report

Report on the field test carried out at HPI of the initial integrated system, the updated integrated system, and integration into VideoLectures.Net platform.

System prototype version 3

The final TraMOOC prototype that integrates the final translation models, taking into account end user and domain expert evaluation, as well as the premium add-on services offered.

Translation prototype version 2

Intermediate machine translation systems for all language pairs.

Translation prototype version 3

Final machine translation system for all language pairs

System prototype version 1

Initial prototype of the TraMOOC system that integrates the initial translation models.

Translation prototype version 1

Initial machine translation systems for 4 language pairs

System prototype version 2

The second TraMOOC prototype that integrates refined translation models for all eleven languages into the open-source web service.

Public website

The TraMOOC public website will provide information about the project its objectives challenges and expected results In addition it will be regularly updated with the projects progress and achievements Finally it will provide detailed information upon the services offered by the translation platform

Data management plan

The Data Management Plan describes the data management life cycle for all data sets that will be collected processed or generated by the research project

Publications

TraMOOC Project overview

Auteurs: Valia Kordoni (UBER)
Publié dans: Proceedings of the 18th Annual Conference of the European Association for Machine Translation (EAMT2015), Numéro May 2015, 2015, Page(s) 217
Éditeur: European Association for Machine Translation

Mixed-Domain vs. Multi-Domain Statistical Machine Translation

Auteurs: Alexandra Birch (UEDIN)
Publié dans: Proceedings of Machine Translation Summit (MT Summit), Numéro October 2015, 2015
Éditeur: Machine Translation Summit (MT Summit)

A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation

Auteurs: R. Sennrich (UEDIN), Barry Haddow (UEDIN)
Publié dans: Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP 2015), Numéro September 2015, 2015
Éditeur: Conference on Empirical Methods in Natural Language Processing (EMNLP 2015)

The Edinburgh/JHU Phrase-based Machine Translation Systems for WMT 2015

Auteurs: Barry Haddow (UEDIN), Matthias Huck (UEDIN), Alexandra Birch (UEDIN), N. Bogoychev (UEDIN), P. Koehn (UEDIN)
Publié dans: Proceedings of the Tenth Workshop on Statistical Machine Translation, Numéro September 2015, 2015, Page(s) 126-133, ISBN 978-1-941643-32-7
Éditeur: Tenth Workshop on Statistical Machine Translation

Enhancing Access to Online Education: Quality Machine Translation of MOOC Content

Auteurs: Valia Kordoni (UBER), Antal Van Den Bosch (RU), Katia Kermanidis (IURC), Vilelmini Sosoni (IURC), Kostadin Cholakov (UBER), Iris Hendrickx (RU), Matthias Huck (UEDIN)
Publié dans: Proceedings of the 10th edition of the Language Resources and Evaluation Conference, Numéro May 2016, 2016, Page(s) 16-22
Éditeur: 10th edition of the Language Resources and Evaluation Conference

The Edinburgh Machine Translation Systems for IWSLT 2015

Auteurs: Alexandra Birch (UEDIN), Matthias Huck (UEDIN)
Publié dans: Proceedings of the 12th International Workshop on Spoken Language Translation (IWSLT), Numéro December 2015, 2015, Page(s) 31-38
Éditeur: 12th International Workshop on Spoken Language Translation (IWSLT)

chrF: character n-gram F-score for automatic MT evaluation

Auteurs: Maja Popovic (UBER)
Publié dans: Proceedings of the Tenth Workshop on Statistical Machine Translation, Numéro September 2015, 2015, Page(s) 392-395
Éditeur: Association for Computational Linguistics

Asistent – A Machine Translation System for Slovene, Serbian and Croatian

Auteurs: Mihael Arčan (UBER), Maja Popovic (UBER), Paul Buitelaar (UBER)
Publié dans: Proceedings of the 10th conference on Language Technologies and Digital Humanities, Numéro September 2016, 2016, Page(s) 13-20
Éditeur: Slovenian Society for Language Technologies

The University of Edinburgh’s Neural MT Systems for WMT17

Auteurs: R. Sennrich (UEDIN)
Publié dans: Proceedings of the EMNLP 2017 Second Conference on Machine Translation (WMT17), Numéro September 2017, 2017, Page(s) 1-11
Éditeur: EMNLP 2017 Second Conference on Machine Translation (WMT17)

Nematus: a Toolkit for Neural Machine Translation

Auteurs: R. Sennrich (UEDIN), Alexandra Birch (UEDIN), Barry Haddow (UEDIN)
Publié dans: Proceedings of the 15th Conference of the European Association for Computational Linguistics (EACL 2017), Numéro April 2017, 2017, Page(s) 1-4
Éditeur: European Association for Computational Linguistics

Enlarging scarce in-domain English-Croatian corpus for SMT of MOOCs using Serbian

Auteurs: Maja Popovic (UBER), Kostadin Cholakov (UBER), Valia Kordoni (UBER), Nikola Ljubešić (UBER)
Publié dans: Proceedings of the 26th International Conference on Computational Linguistics, Numéro December 2016, 2016, Page(s) 97-105
Éditeur: 26th International Conference on Computational Linguistics

A Comparative Quality Evaluation of PBSMT and NMT using Professional Translators

Auteurs: Sheila Castilho (DCU), Joss Moorkens (DCU), Federico Gaspari (DCU), R. Sennrich (UEDIN), Vilelmini Sosoni (IURC), Panayota Georgakopoulou (DME), Pintu Lohar (DCU), Andy Way (DCU), Antonio Miceli Barone (UEDIN), Maria Gialama (DME)
Publié dans: Proceedings of the 16th Machine Translation Summit -2017 Vol. 1 Research Track, Numéro September 2017, 2017, Page(s) 16-131
Éditeur: Asia-Pacific Association for Machine Translation (AAMT)

Language Related Issues for Machine Translation between Closely Related South Slavic Languages

Auteurs: Maja Popovic (UBER), Mihael Arčan (UBER), Filip Klubička
Publié dans: Proceedings of the Third Workshop on NLP for similar Languages, Varieties and Dialects (VarDial3), Numéro December 2016, 2016, Page(s) 43-52
Éditeur: Third Workshop on NLP for similar languages, varieties and dialects

Regularization techniques for fine-tuning in neural machine translation

Auteurs: R. Sennrich (UEDIN), Antonio Miceli Barone (UEDIN), Barry Haddow (UEDIN), Ulrich Germann (UEDIN)
Publié dans: Proceedings of the EMNLP 2017 Second Conference on Machine Translation (WMT17), Numéro September 2017, 2017, Page(s) 1489-1494
Éditeur: EMNLP 2017 Second Conference on Machine Translation (WMT17)

Comparing Language Related Issues for NMT and PBMT between German and English

Auteurs: Maja Popovic (UBER)
Publié dans: Proceedings of the 20th Annual Conference of the European Association for Machine Translation (EAMT2017) (The Prague Bulletin of Mathematical Linguistics 108, June 2017), Numéro May 2017, 2017, Page(s) 209-220, ISSN 0032-6585
Éditeur: 20th Annual Conference of the European Association for Machine Translation (EAMT2017) Organizing Committee

INTEGRATING MACHINE TRANSLATION INTO MOOCS

Auteurs: Sheila Castilho, Federico Gaspari, Joss Moorkens, Andy Way
Publié dans: EDULEARN17 Proceedings, Numéro July 2017, 2017, Page(s) 9360-9365, ISBN 978-84-697-3777-4
Éditeur: IATED
DOI: 10.21125/edulearn.2017.0765

PE2rr Corpus: Manual Error Annotation of Automatically Pre-annotated MT Post-edits

Auteurs: Maja Popovic (UBER), Mihael Arčan (UBER)
Publié dans: Proceedings of the 10th edition of the Language Resources and Evaluation Conference, Numéro May 2016, 2016, Page(s) 27-32
Éditeur: 10th edition of the Language Resources and Evaluation Conference

chrF deconstructed: beta parameters and n-gram weights

Auteurs: M. Popovic
Publié dans: Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016), Numéro August 2016, 2016, Page(s) 499-504
Éditeur: Association for Computational Linguistics

The AMU-UEDIN Submission to the WMT16 News Translation Task: Attention-based NMT Models as Feature Functions in Phrase-based SMT

Auteurs: R. Sennrich
Publié dans: Proceedings of the ACL 2016 First Conference on Machine Translation (WMT16), Numéro August 2015, 2016, Page(s) 319-325
Éditeur: Association for Computational Linguistics

Edinburgh Neural Machine Translation Systems for WMT 16

Auteurs: R. Sennrich, B. Haddow and A. Birch
Publié dans: Proceedings of the ACL 2016 First Conference on Machine Translation (WMT16), Numéro August 2016, 2016, Page(s) 371-376
Éditeur: Association for Computational Linguistics

Edinburgh’s Statistical Machine Translation Systems for WMT16

Auteurs: R. Sennrich, B. Haddow and M. Huck
Publié dans: Proceedings of the ACL 2016 First Conference on Machine Translation (WMT16), Numéro August 2016, 2016, Page(s) 399-410
Éditeur: Association for Computational Linguistics

Casting some Light on Experts’ Experience with Translation Crowdsourcing

Auteurs: Vilelmini Sosoni (IURC)
Publié dans: Journal of Specialised Translation, Numéro 28, Numéro July 2017, 2017, Page(s) 362 - 384, ISSN 1740-357X
Éditeur: University of Roehampton

Is Neural Machine Translation the New State of the Art?

Auteurs: Sheila Castilho, Joss Moorkens, Federico Gaspari, Iacer Calixto, John Tinsley, Andy Way
Publié dans: The Prague Bulletin of Mathematical Linguistics, Numéro 108/1, 2017, Page(s) 109-120, ISSN 1804-0462
Éditeur: 20th Annual Conference of the European Association for Machine Translation (EAMT2017) Organizing Committee
DOI: 10.1515/pralin-2017-0013

Potential and Limits of Using Post-edits as Reference Translations for MT Evaluation

Auteurs: Maja Popovic (UBER), Mihael Arčan (UBER), Arle Lommel (UBER)
Publié dans: Baltic Journal of Modern Computing (Proceedings of the 19th Annual Conference of the European Association of Machine Translation), Numéro Vol. 4, 2016, Page(s) 218-229, ISSN 2255-8950
Éditeur: University of Latvia

Livrables

Publications

Partager cette page

Télécharger