Skip to main content

Computational SLA

Computational SLA* (CompSLA for short) is an initiative for organizing a series of Shared Tasks with reference to second language (L2) learning. Its aim is to encourage languages that do not have extensive collections of  annotated learner data to take a step towards sharing that data with wider community and stimulating research around it.

The initiative was taken during a NLP4CALL workshop in 2021 (see slides).

*SLA = Second Language Acquisition

 

First CompSLA task

Currently, a group of researchers representing L2 datasets for Czech, English, German, Italian and Swedish are working on the preparation of the first shared task.

Venue

The intended venue is NLP4CALL-2023, co-located with the Nodalida conference in Faeroe Islands.

Timeline (preliminary)

  • 09 December 2022 - first call for participation. Training data released, CodaLab opens for team registrations.
  • 08 January 2023 - validation server released online
  • 14 February 2023 - test data released
  • 21 February 2023 - system submission deadline (system output, models, code, fact sheets, extra data, etc)
  • 28 February 2023 - results announced
  • 28 March 2023 - paper submission deadline (system descriptions)
  • 14 April 2023 - paper reviews sent to the authors
  • 05 May 2023 - camera-ready deadline
  • 22(-24) May 2023 - presentations of the systems at workshop NLP4CALL

Task definition (preliminary)

Multi-Class Grammatical Error Detection (MuClaGED)

Details to follow

Targeted languages and datasets

Organizers

Contact: Elena Volodina < elena . volodina @ svenska . gu . se >