PostEd

Louise Brunette, UQO (R & D)

Alain Désilets, NRC (R & D)

PostEd is a research group that aims to improve the professional practice of post-editing by humans. Post-editing is defined as human intervention on original or machine-translated texts.

The research methodology adopted for this project involves observing post-editors at work in a professional environment to better understand this relatively new practice. PostEd objectives are to facilitate the post-editing process and implement more effective work methods.

For more information:

Louise.Brunette@uqo.ca
Alain.Desilets@nrc-cnrc.gc.ca

Automatic document classification and categorization

Cyril Goutte, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The Automatic document classification and categorization project aims to develop methods for automatically processing documents, particularly textual documents, with two goals:

• -Classification (clustering): when someone wants to determine which groupings of documents in a collection would be relevant for classification purposes;
  
• -Categorization: when someone wants to place an uncategorized document into a previously established classification system.

For more information:

Cyril.Goutte@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

Archive publishing

(Virtual archives)

Marie Grégoire, UQO (R & D and project development)

Also:

Pierre Isabelle, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the Archive publishing project is to make accessible through digitization and publication archived documents that provide information about Nouvelle France and First Nations societies, among others.

These documents are being published in two stages:

For the first stage, archivist-paleographers are transcribing the texts. The LTRC´s ultimate goal is to progress from manual transcription (current practice) to semi-automatic transcription (medium term). Transcription makes the documents accessible to researchers and students and will serve as the basis for new studies.

The second stage is to digitally index the information in the transcribed texts, as practiced, for instance, by the ATILF laboratory (www.atilf.fr/), which is affiliated with the NancyUniversity. This second stage will allow semantic dictionaries and indexes to be created. Moreover, the evolution of ideas, feelings, figures, impressions of places, etc. will be easier to study and understand.

http://www.archivesvirtuelles.com/

For more information:

Marie.Gregoire@uqo.ca
Pierre.Isabelle@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

(see the project summary in PDF format)

Christiane Melançon, UQO (R & D)

Alain Désilets, NRC (R & D)

The Observation of Technologized Language Practices (OTLP-1) project aims to understand language professional/machine interaction by studying human translation in an IT environment (bitexts, translation-memory-type tools and terminology databases).

For more information:

Christiane.Melancon@uqo.ca
Alain.Desilets@cnrc-nrc.gc.ca

PORTAGE

(see the project summary in PDF format)

Roland Kuhn, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the PORTAGE project is to develop technology that will allow a computer to translate a text from one language to another and roughly assess translation quality. PORTAGE uses a statistical approach to machine translation, based on sentence segments (‘phrase-based’).

The word-alignment technology (PORTalign) derived from PORTAGE is available under commercial licence.

http://iit-iti.nrc-cnrc.gc.ca/projects-projets/portage_e.html

For more information:

Roland.Kuhn@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

PORTAGE : GALE & SMART

See PORTAGE project (above).

See PORTAGE project (above). There is no specific description for these projects.

(see the project summary in PDF format)

Caroline Barrière, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

TerminoWeb is a flexible software environment designed for terminologists but also useful for experts (terminology search in a specialized field) or even learners (acquiring specialized vocabulary in learner-selected fields). TerminoWeb improves the efficiency of thematic search by allowing (1) the creation of a specialized corpus of documents from the Web, (2) terminology extraction from the corpus, and (3) semantic exploration of the corpus.

For the TerminoWeb project, please see:

http://iit-iti.nrc-cnrc.gc.ca/projects-projets/terminoweb_e.html

A TerminoWeb prototype is accessible online at:

http://termino.iit.nrc.ca/

TerminoWeb also acts as an R & D environment at the NRC, in the field of computational terminology.

For more information:

Caroline.Barriere@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

Alain Désilets, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the Collaborative translation and terminology project is to develop tools and processes to help collaborative translation and terminology practices, in a Wikipedia-like way.

For more information:

Alain.Desilets@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

TransCheck

(see the project summary in PDF format)

Elliott Macklovitch, UdeMontréal (R & D)

Pierre Isabelle, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

TransCheck is an automatic translation checker initially developed at the Université de Montréal´s RALI lab. Development is now continuing jointly by RALI and the NRC.

TransCheck reads a text and its translation and then identifies pre-defined translation errors, such as terminological inconsistency, incorrect terminology (particularly false cognates), paralinguistic expressions (particularly numeric expressions), proper nouns, omissions and additions. TransCheck can also be adapted, upon request, to pre-assess terminological consistency for a specific group of users.

TransCheck has been developed to the point where it could be used at a translation company, and trials to this effect are being conducted.

For more information:

Pierre.Isabelle@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

WeBiNuk

Benoit Farley / Alain Désilets, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the WeBiNuk project is to develop and assess an English-Inuktitut translation memory created from multilingual parallel content found on the Web.

The project is one of the IIT´s Uqausiit activities. This group is developing IT tools to process Inuktitut.

http://iit-iti.nrc-cnrc.gc.ca/projects-projets/uqausiit_e.html

For more information:

Benoit.Farley@cnrc-nrc.gc.ca
Alain.Desilets@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

WeBiText

A. Désilets / Benoit Farley, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the WeBiText project is to develop and assess a very large, multilingual, multi-field translation memory built from parallel multilingual content from the Web.

Alain.Desilets@cnrc-nrc.gc.ca
Benoit.Farley@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

L2Corr

A. Désilets, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The goal of the L2Corr project is to develop and assess text correction software targeted at second-language writers. This system will correct mistakes that are commonly made by second-language writers but which are usually ignored by standard grammar and spelling checkers. These errors include preposition choice ("I grew up *at* Montreal").

For more information:

Alain.Desilets@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

Barçah

(see the project summary in PDF format)

Jean Quirion, UQO (Terminometry and Barçah testing)

Also:

Michel Mellinger, NRC (Transfer of Barçah software developed by the NRC)

The goal of the Barçah project was to develop software ("Barçah") that would automate comparative and diachronic terminometry to help facilitate the analysis of terminology and large collections of documents.

For more information:

Michel.Mellinger@cnrc-nrc.gc.ca

BiTerm

Michel Simard, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

TransSearch is a service offered online by Terminotix (RALI technology developed by Michel Simard, now with the ILT Group).

The goal of the BiTerm project was to further develop such a concordancer at the ILT Group and to use a prototype to illustrate improved functionalities compared to those of TransSearch.

For more information:

Michel.Simard@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

Multilingual Chat

The goal of the Multilingual Chat project was to develop a multilingual chat system that would allow multiple users to chat simultaneously via a computer network (usually the Internet) in the language of their choice.

For more information:

Alain.Charbonneau@uqo.ca
Roger.J.Racine@tpsgc.gc.ca

ILO: Bilingual research for the diffusion of scientific and technical documents from federal departments and agencies

Caroline Barrière, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The ILO project focused on translingual documentary research and particularly on how accessible scientific and technical information generated by the Government of Canada is to Canadian citizens in both official languages.

The ultimate goal of this project was to make available to the public a system for conducting bilingual document searches on the Internet. The initial feasibility stage was concluded and provided positive results.

An acronym expansion software module (ACRO) was developed by this project and is available under licence.

For more information:

Caroline.Barriere@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

Revision memory

Roger Racine, BtB / Pierre Isabelle, NRC (R & D)

Also:

Michel Mellinger, NRC (collaborations and transfer)

The Revision memory project (MemRev) aimed to explore a new type of computer-assisted translation and writing tool. This system would be able to suggest local changes to a draft text to help improve its quality.

For more information:

Roger.J.Racine@tpsgc.gc.ca
Pierre.Isabelle@cnrc-nrc.gc.ca
Michel.Mellinger@cnrc-nrc.gc.ca

PAT – Collecting translation dictations

A. Désilets, NRC (R & D)

The goal of the PAT project was to assess commercial voice-recognition technology when used for translation dictation purposes. The project aimed to assess the impact of this technology on translator productivity and the possibility of improving recognition rates by combining it with machine translation systems.

For more information:

Alain.Desilets@cnrc-nrc.gc.ca

SISTIL
(Integrated system for providing technological support to the language industry)

The goal of the SISTIL project was to develop an architecture model, a structure and a prototype for integrating translation tools, systems and processes.

For more information:

jacques_bedard@videotron.ca