Papers by Mihai Vlad Constantinescu
RED v2: Enhancing RED Dataset for Multi-Label Emotion Detection (2022.lrec-1)
Copied to clipboard
| Challenge: | RED is a machine learning-based resource developed for the automatic detection of emotions in Romanian texts. |
| Approach: | They propose an open-source extension of RED by adding trust and surprise . they propose two variants of ground truth suitable for multi-label classification and text regression . |
| Outcome: | The proposed model is based on two models with two transformer models, the Romanian BERT and the multilingual XLM-Roberta model, in categorical and regression settings. |