Maximizing SER Performance with ICR-SEA: A Novel Framework for Grid-Search Optimization
Bibliographic record
Abstract
This proposed work develops a novel strategy to optimize CNN-LSTM (Convolutional Neural Network-Long Short-Term Memory for SER (Speech Emotion Recognition) using Mel-spectrograms as input features with the RAVDESS (Ryerson Audio-Visual Database of Emotional Speech and Song) data corpus. Data pre-preparation, including dataset class balancing and augmentation improves model generalization. The hyperparameter tuning strategy employs the Iterative Candidate Refinement with Successive Epoch Adjustment (ICR-SEA) algorithm, which efficiently tunes the CNN-LSTM architecture and training parameters. This approach systematically filters best candidates while dynamically adjusting the number of training epochs, ensuring computational effectiveness. This hyperparameter tuning process investigates changes in convolutional filters, kernel sizes, dense layer sizes, dropout rates, batch sizes, learning rates, and epochs. The top-performing candidate model is precisely evaluated and filtered for the next iteration. The optimized model reaches a classification accuracy of 85.36% approximately during hyperparameter tuning. This research emphasizes the effectiveness of the ICR-SEA algorithm in hyperparameter optimization and the significance of data augmentation in improving the generalizability and performance of emotion recognition systems, setting a benchmark for future work with the RAVDESS dataset.
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 | 0.002 | 0.005 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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".