We would like to extend an invitation to participate in the shared task on Formality Control for Spoken Language Translation organized at IWSLT 2022. This shared task invites participants to advance research on controlling formality in machine translation for a diverse set of languages with a focus on zero-shot learning. To advance research in this area, the organizers have provided an annotated dataset and a targeted evaluation metric.
Important links:
- Formality task webpage: https://iwslt.org/2022/formality
- Data, metric and per-trained models: https://github.com/amazon-research/contrastive-controlled-mt
- Google Group: https://groups.google.com/g/iwslt-evaluation-campaign
- IWSLT conference webpage: https://iwslt.org/2022/
HOW TO PARTICIPATE
Please join the IWSLT Evaluation Campaign Google Group and access the registration form shared on this group.
We have released formality-annotated training and dev data and an automatic metric here: https://github.com/amazon-research/contrastive-controlled-mt. Allowable generic training data is listed on the task website. Task submissions will be made on a dedicated github repository. More information will be added here at a later date.
More resources:
- On February 1st, we will release pre-trained Sockeye 3 models and scripts that participants can use as a basis for their models
IMPORTANT DATES
- January 15, 2022: training data released
- February 1, 2022: pre-trained models released
- March 14, 2022: evaluation data released
- March 24, 2022: translation submission deadline
- March 25, 2022: reference translations released
- April 1, 2022: system paper submission deadline
- April 22, 2022: paper notification
- May 1, 2022: camera ready deadline
- May 26-27, 2022: IWSLT conference (Dublin, Ireland and virtual)
LANGUAGE PAIRS
- English → German
- English → Hindi
- English → Japanese
- English → Spanish
Zero-shot languages (no training data released):
- English → Italian
- English → Russian
You may participate in any subset of the language pairs.
ORGANIZING COMMITTEE
Maria Nadejde, Amazon AWS Translate
Anna Currey, Amazon AWS Translate
Benjamin Hsu, Amazon AWS Translate
Xing Niu, Amazon AWS Translate
Georgiana Dinu, Amazon AWS Translate
Marcello Federico, Amazon AWS Translate