Emotion Detection in Text using Mixup-Augmented BERT Representations
Bibliographic record
Abstract
Emotion recognition in text is gaining significant applicative interest because of its potential applications in health care, social media analytics, and improving customer experiences. This work presented a transformer-based approach where BERT was combined with the mixup data augmentation technique, boosting human emotion classification. In this paper, the authors use the GoEmotions dataset comprising 58,000 Reddit comments labeled into 27 classes of emotions. Therefore, it can become a very challenging benchmark, due to class imbalance and semantic overlap. Preprocessing steps include tokenization, normalization, and encoding of labels. The embedding is done using BERT. Mixup, at the embedding level, changes how phrases and their labels are presented. The model then learns much knowledge from these fake examples. The BERT model that has been augmented using Mixup outperforms the BERT classifier that does not alter the data. The F1-score is 94.3%, the recall is 92.1%, and the precision is 98.0%. The results show an improvements in minority classes such as pride and grief, indicating the effectiveness of mixup in addressing imbalance. This methodology provides a lightweight, scalable, and robust solutions in contrast to current solutions that depend on alterations in the architectural or affect-enriched embeddings. These results led to the development of mixup-augmented transformers, which show promise for recognizing emotions in real-world situations like monitoring mental health and engaging users in personalized manner.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".