Multidimensional Affective Analysis for Guarani/Jopara: emotion, humor and offense.
Affect dimensions (tweets primarily written in Guarani):
- Emotion-recognition: https://huggingface.co/datasets/mmaguero/gn-emotion-recognition
- Humor-detection: https://huggingface.co/datasets/mmaguero/gn-humor-detection
- Offensive-language-identification: https://huggingface.co/datasets/mmaguero/gn-offensive-language-identification
- From scratch (trained on Guarani Wiki data):
- Continuously pre-trained (or pre-fine-tuned, also trained on Wiki data):
- Bert-based models (fine-tuned with our above base models): https://colab.research.google.com/drive/1cjtykOqGz7B74yX452k5T0MbZmjtdpYc
- NCRFpp-based models (trained with https://github.com/jiesutd/NCRFpp library): https://colab.research.google.com/drive/1kzdxn0cdg7_bp6hSpAx3I8Pw9vul-8IM
@article{aguero-et-al2023multi-affect-low-langs-grn,
title={Multidimensional Affective Analysis for Low-resource Languages: A Use Case with Guarani-Spanish Code-switching Language},
author={Agüero-Torales, Marvin Matías, López-Herrera, Antonio Gabriel, and Vilares, David},
journal={Cognitive Computation},
year={2023},
publisher={Springer},
notes={https://link.springer.com/article/10.1007/s12559-023-10165-0#citeas}
}