This paper has been retracted. The evaluation in the paper has been performed with an erronous evaluation script. The test data in the shared task has a considerable overlap with the training data. Because of both issues, the performance estimates of the models are too optimistic.
Abstract
We present the results of the WASSA 2023 Shared-Task 2: Emotion Classification on code-mixed text messages (Roman Urdu + English), which included two tracks for emotion classification: multi-label and multi-class. The participants were provided with a dataset of code-mixed SMS messages in English and Roman Urdu labeled with 12 emotions for both tracks. A total of 5 teams (19 team members) participated in the shared task. We summarized the methods, resources, and tools used by the participating teams. We also made the data freely available for further improvements to the task.- Anthology ID:
- 2023.wassa-1.56
- Original:
- 2023.wassa-1.56v1
- Version 2:
- 2023.wassa-1.56v2
- Version 3:
- 2023.wassa-1.56v3
- Volume:
- Proceedings of the 13th Workshop on Computational Approaches to Subjectivity, Sentiment, & Social Media Analysis
- Month:
- July
- Year:
- 2023
- Address:
- Toronto, Canada
- Editors:
- Jeremy Barnes, Orphée De Clercq, Roman Klinger
- Venue:
- WASSA
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 587–595
- Language:
- URL:
- https://aclanthology.org/2023.wassa-1.56
- DOI:
- 10.18653/v1/2023.wassa-1.56
- PDF:
- https://preview.aclanthology.org/dois-2013-emnlp/2023.wassa-1.56.pdf