A Novel Approach to Speech Emotion Recognition Using Wav2Vec2
Bibliographic record
Abstract
Speech Emotion Recognition (SER), which allows computers to precisely understand and react to human emotions, is essential for improving human-computer interaction. This subject is crucial because it has the potential to significantly improve the user experience in applications including education and e-Iearning, mental health, customer service, Automotive Industry, and others. Speech Emotion Recognition (SER) systems that used classic machine learning techniques usually extract handcrafted audio features such as pitch, loudness, energy, spectral flux, and others each of which captured a distinct part of the emotional tone in speech. Despite their ease of use and interpretability, traditional machine learning models such as Support Vector Machines (SVMs) and Decision Trees frequently had trouble handling complex emotional patterns and speaker variances. These limitations led researchers to use deep learning models, specifically Recurrent Neural Networks (RNNs), which could analyze speech as a sequence and capture some temporal relationships. By simulating the evolution of emotions over time, RNN s improved emotion identification. However, RNNs faced their own limitations, particularly in handling long-range dependencies and in maintaining performance when faced with diverse emotional expressions or noisy environments. These challenges led to the Transformer models. Unlike previous models, Transformers can capture long-range dependencies and complex patterns in data, making them well-suited for capturing the nuanced expressions necessary for accurate emotion detection. Therefore, in this research, we will present a Wav2Vec2 transformer SER system that has been optimized on emotional speech data to recognize a variety of emotional states. We conducted our experiment on the small dataset called Ryerson Audio-Visual Database of Emotional Speech and Song (RA VDESS) and achieved an improvement in accuracy and robustness over previous methods. It achieved a testing accuracy of 89% when compared to SVM using GWO optimizer, which achieved an accuracy of 76%, and the LSTM using GWO optimizer, which achieved an accuracy of 71 %.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.003 |
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 source (direct Gemma or distilled Codex), 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".