Skip to content

Code and issue tracker for WMT21 source-based DA campaigns

License

Notifications You must be signed in to change notification settings

AppraiseDev/WMT21SrcDA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

WMT21SrcDA

Code and issue tracker for WMT21 source-based DA campaigns

Updates

  • 2021-08-20: Accounts are live now, please contact {tomkocmi,chrife,rogrundk}@microsoft.com about your accounts
  • 2021-08-20: WMT21.appraise.cf goes live

TL;DR

  • WMT21 will feature source-based direct assessment (DA) with document-level annotation
  • We will focus on non-English target languages, and annotators need to be native speakers of the target language
  • Teams receive a number of anonymous accounts to perform annotations for language pairs of their choice
  • The number of accounts received is based on the number of team's primary submissions
  • Please use the Github issue tracker to report any problems

Timeline

  • 2021-08-06: research teams request for annotation languages
  • 2021-08-20: accounts shared with research teams
  • 2021-08-20: annotation starts
  • 2021-09-16: annotation ends

Campaign overview

How will the annotation work?

Annotations will be collected in Appraise, implementing document-level, source-based direct assessment. For every language pair, we will generate anonymised accounts which are pre-assigned to exactly two annotation tasks.

How much annotation work is needed?

One account contain two tasks, each involving 100 sentence-level judgements. The average annotation time for one task is ca. 45 minutes, so we assume that each account is 90 minutes of annotation work.

Each research team is expected to contribute ten hours of annotation work per primary system submission. We assign you with only nine hours of annotations per primary system to keep buffer in case the annotations would take longer. This translates to 12 completed tasks or 6 completed accounts in total, on Appraise, per primary submission.

Who can be an annotator?

The source-based evaluation requires native speakers of the non-English target language who are also proficient in English and can assess translation quality from English into their native language.

How are accounts distributed?

Account distribution is based on first-come, first-served basis, i.e., once an account is marked as ``assigned'' to a team you cannot claim it anymore. Please note that you can provide annotations for any language pair, not only language pairs you have submitted primary systems to.

Each research team should designate a team leader who is then responsible to reserve the required amount of accounts for their team.

Where can I claim my account?

Please contact {tomkocmi,chrife,rogrundk}@microsoft.com to request your accounts.

How can I sign into Appraise?

For each account, we provide a single sign-on (SSO) URL. This allows you to sign into Appraise with a single click on the URL, making access very easy.

What information do you store about me?

There is no personal information attached to the annotation accounts. We capture your assessments and related metadata, such as annotation start and end time as well as duration per single assessment.

How does source-based DA work?

This year the campaign features a document-level annotation view with the entire document presented on a single screen in two columns. The left column has source sentences and the right column has corresponding candidate translations (see the screen), that you will be scoring using a slider.

Screenshot of WMT21

The annotation process consists of scoring individual sentences one-by-one and then assigning a single score for the entire document. Assigning the document-level score (the slider at the very bottom of the page) becomes available after all previous sentences are scored.

Submitting a score will automatically move the annotation to the next sentence and open a new slider. A dot on the right side of the translated sentence roughly indicates the assigned score: a reddish dot means a lower score (worse translation), and a greenish dot means a higher score (better translation). The tick means that the score has been successfully collected by the server and the progress is saved (i.e. the web page can be reloaded without loosing the progress made on the current document). If you decide to change the assigned score, it is possible to do so at any point of the annotation process by clicking on the sentence text to expand a slider, updating the score, and clicking 'Update'. When the document-level score is submitted, the annotation process continues with the next document until all documents assigned to your account are annotated.

How can I report problems?

Please use the Github issue tracker to report any problems. You can also contact us via {tomkocmi | chrife | rogrundk} [at] microsoft [dot] com.

Frequently asked questions

Are there any reference based annotations or into English translations?

No, this campaign is source based DA for non-english target language only.

Do I need to annotate languages of my primary system submissions?

No, that is not necessary. We balance annotations accross teams that speak various languages. You can annotate any language pair in which you are proficient.

Why does some documents in annotation are same/similar?

Various systems have translated the same document, therefore similar translations are annotated. Furthermore, it may happen that the same document occurs in multiple tasks.

We completed a task in each account but are asked to do more?

That is likely because you have finished only one task for some accounts when we require completing both tasks assigned to each account. If the account is fully completed, then you will not see any link on the Appraise dashboard.

About

Code and issue tracker for WMT21 source-based DA campaigns

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published