Shared Task Hackathon on NLP

We are pleased to announce the Shared Task Hackathon on natural language processing, which will take place in December 21, 2021. It will be collocated with iSAI-NLP 2021. The shared task hackathon brings together researchers, professors, students, NLP engineers, and NLP developers in a Kaggle-style competition for Thai language processing. The aim of this hackathon is to tackle four common tasks: word segmentation, POS tagging, named entity recognition, and sentence segmentation.

The competition is open and inclusive for Thai and non-Thai contestants. All contestants can register as a team of up to three (3) persons. The event is organized in English, and all distributed documents will be in English. The contestants can submit necessary documents and give a pitching talk either in English or in Thai.

Three prices will be given to the winners.

  • Winner 10,000 THB
  • 1st Runner-Up 8,000 THB
  • 2nd Runner-Up 5,000 THB
  • 3rd & 4th Runners-Up 2,500 THB

All the winners will also be given free annual membership of AIAT.

The registration is now open.

Registration Click Here!!


Stage 1: Data Distribution
On December 4, 2021 00:00 GMT+07, we will distribute the dataset of more than 50 million words that complies with the LST20 annotation guideline. This dataset is annotated with word boundaries, POS tags, named entity labels, and sentence boundaries. All teams are encouraged to develop efficient NLP models for the four tasks based on the dataset and leverage additional language models e.g. Word2Vec and BERT.

Stage 2: Result Submission
  • On December 14, 2021, we will distribute an unlabeled test data to all teams for evaluation. The contestants are required to annotate the four layers of linguistic information to this test data with their models.
  • Each team will be evaluated and ranked with respect to the average F1 accuracies of all tasks. All teams can submit the annotated test data for evaluation for up to 10 times. The submission will be closed on December 17, 2021 at 23:59 GMT+07.
  • Top-5 teams will be selected for the pitching stage.

Stage 3: Technical Pitching
After the submission is closed, the selected teams will be notified for the technical pitching session. Each team can choose to give a pitching talk in English and Thai. Each team will be given seven (7) minutes for pitching and five (5) minutes for question answering.

Each team will be equally assessed with the following criteria.

  • Problem statement
  • Solution
  • Technical novelty & differentiation
  • Presentation skills

The competition results and each team's technical details will be published in the JIIST Journal.