Translation Quality Webinar: Analyzing Translation Quality

11 October, 2017, 05:00 - 06:00 pm CEST

Overview

This bimonthly webinar is open to buyers and vendors of translation and localization services interested in translation quality. In general two invited speakers present their use case or introduce a topic. After the presentations, an expert panel asks questions and the audience can raise questions that are answered by the presenters.

Translation quality evaluation is problematic. In 2011, TAUS conducted a survey among its members. We found that despite very detailed and strict error-based evaluation models the satisfaction levels with both translation quality and the evaluation process itself were very low. QE models are static, that is, there is a ‘one size fits all’ approach. Little consideration is given to multiple variables such as content type, communications function, end user requirements, context, perishability, or mode of translation generation.

Agenda

Analyzing Translation Quality to help Improve Machine Translation

In this webinar we will present the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. After an introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. Further, we will describe the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment. Finally, we will present how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.

Agenda

  1. Welcome by Jaap van der Meer
  2. Introduction to the QT21 project by Christian Dugast
    Presenting the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. 
  3. Different approaches, methods, and tools to analyze quality by Lucia Specia
    An introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. 
  4. Annotated data generated in the project by Aljoscha Burchardt
    Discussing the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment.
  5. Continuous learning: incorporating QT21 data in an Automatic Post-editing system by Marco Turchi
    Presenting how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.
  6. Q&A

Speakers

Aljoscha Burchardt | DFKI

Aljoscha Burchardt is lab manager at the Language Technology Lab of the German Research Center for Artificial Intelligence (DFKI GmbH). His interests include the evaluation of (machine) translation quality and the inclusion of language professionals in the MT R&D workflow. Burchardt is co-developer of the MQM framework for measuring translation quality. He has a background in semantic Language Technology.


Christian Dugast | DFKI

Christian received his Ph.D. degree in Computer Science from the University of Toulouse (France) in 1987. He started his career as a research scientist in Automatic Speech Recognition (ASR) at Philips Research Laboratories under Hermann Ney. In 1998, he entered the business world and build up from scratch the European subsidiary of Nuance Communications. In 2002, Christian co-founded VoiceObjects, a platform needed for the industrialisation of voice services. Since 2007 Christian is consultant for Natural Language Processing technologies helping start-ups. Always with a foot into research, he is currently managing QT21, an European Commission funded Research & Innovation Action (RIA) project in which 10 leading academic European research labs and three industrial partners are looking at improving Machine Translation for morphologically rich languages.


Lucia Specia | University of Sheffield

Dr. Lucia Specia is Professor of Language Engineering at the Department of Computer Science of the University of Sheffield. Her research focuses on various aspects of data-driven approaches to multilingual language processing, with applications to Machine Translation, Quality Estimation and Text Adaptation. She is the recipient of an ERC Starting Grant on Multimodal Machine Translation (2016-2021) and is involved in other funded research projects on Machine Translation (QT21 21 and CRACKER) and Text Adaptation (SIMPATICO). Before joining the University of Sheffield in 2012, she was Senior Lecturer at the University of Wolverhampton, UK (2010-2011), and research engineer at the Xerox Research Centre, France (2008-2009). She received a PhD in Computer Science from the University of São Paulo, Brazil, in 2008. She has published over 100 research papers in peer-reviewed journals and conference proceedings. She has served as area and program chair, and on programme committees of numerous leading international conferences and journals, and organised a number of workshops and shared tasks in the area of NLP.


Marco Turchi | FBK

Marco is a researcher in the Human Language Technology Machine Translation (HLT-MT) group at Fondazione Bruno Kessler (FBK) in Trento, Italy. Before joining FBK, he worked as research engineer at the European Commission Joint Research Centre in Italy, at the University of Bristol and at the Xerox Research Centre Europe. He received his Ph.D. degree in Computer Science from the University of Siena, Italy in 2006. His current research is centered around applying machine learning techniques to MT, with particular emphasis on exploiting post-edited data to improve MT quality. He is involved in various funded research projects, including the European initiatives QT21 (Quality Translation 21) and MMT (Modern Machine Translation). He has co-authored more than 80 peer-reviewed scientific publications.


Event Properties

Event Date 11-10-2017 5:00 pm
Event End Date 11-10-2017 6:00 pm
Capacity Unlimited
Individual Price Free
Created By Anne-Maj van der Meer
Registration link https://attendee.gotowebinar.com/rt/8526097327405857026
Overview

This bimonthly webinar is open to buyers and vendors of translation and localization services interested in translation quality. In general two invited speakers present their use case or introduce a topic. After the presentations, an expert panel asks questions and the audience can raise questions that are answered by the presenters.

Translation quality evaluation is problematic. In 2011, TAUS conducted a survey among its members. We found that despite very detailed and strict error-based evaluation models the satisfaction levels with both translation quality and the evaluation process itself were very low. QE models are static, that is, there is a ‘one size fits all’ approach. Little consideration is given to multiple variables such as content type, communications function, end user requirements, context, perishability, or mode of translation generation.

Agenda

Analyzing Translation Quality to help Improve Machine Translation

In this webinar we will present the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. After an introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. Further, we will describe the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment. Finally, we will present how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.

Agenda

  1. Welcome by Jaap van der Meer
  2. Introduction to the QT21 project by Christian Dugast
    Presenting the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. 
  3. Different approaches, methods, and tools to analyze quality by Lucia Specia
    An introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. 
  4. Annotated data generated in the project by Aljoscha Burchardt
    Discussing the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment.
  5. Continuous learning: incorporating QT21 data in an Automatic Post-editing system by Marco Turchi
    Presenting how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.
  6. Q&A
Speakers (9829, 16570, 19601, 16708)

Event Properties

Event Date 11-10-2017 5:00 pm
Event End Date 11-10-2017 6:00 pm
Capacity Unlimited
Individual Price Free
Created By Anne-Maj van der Meer
Registration link https://attendee.gotowebinar.com/rt/8526097327405857026
Overview

This bimonthly webinar is open to buyers and vendors of translation and localization services interested in translation quality. In general two invited speakers present their use case or introduce a topic. After the presentations, an expert panel asks questions and the audience can raise questions that are answered by the presenters.

Translation quality evaluation is problematic. In 2011, TAUS conducted a survey among its members. We found that despite very detailed and strict error-based evaluation models the satisfaction levels with both translation quality and the evaluation process itself were very low. QE models are static, that is, there is a ‘one size fits all’ approach. Little consideration is given to multiple variables such as content type, communications function, end user requirements, context, perishability, or mode of translation generation.

Agenda

Analyzing Translation Quality to help Improve Machine Translation

In this webinar we will present the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. After an introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. Further, we will describe the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment. Finally, we will present how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.

Agenda

  1. Welcome by Jaap van der Meer
  2. Introduction to the QT21 project by Christian Dugast
    Presenting the work done by QT21, a European Commission funded a research project on machine translation that makes use of translation quality analysis in order to improve machine translation. 
  3. Different approaches, methods, and tools to analyze quality by Lucia Specia
    An introduction to the different approaches, methods, and tools that are available to analyze quality (manual and/or automated quality estimations, professional translator based or crowd source based) we will summarize their theoretical pros and cons. 
  4. Annotated data generated in the project by Aljoscha Burchardt
    Discussing the data QT21 has generated on 4 language pairs (typically post editions, error annotations) and analyze their value for quality assessment.
  5. Continuous learning: incorporating QT21 data in an Automatic Post-editing system by Marco Turchi
    Presenting how we could improve machine translation by incorporating this data within an Automatic Post-Editing system what leads to the possibility of building continuously learning machine translation systems.
  6. Q&A
Speakers (9829, 16570, 19601, 16708)
Share this event: