FedAR: Federated Artificial Resampling for Imbalanced Facial Emotion Recognition
Bibliographic record
Abstract
Federated Learning (FL) has emerged as an essential tool for computing devices to participate in collaborative training of deep learning models. However, due to the decentralized distribution of data over clients/local computing devices, the class imbalance problem has become evident, causing severe degradation in the performance of the global model. Motivated by the emergence of FL models in emotion recognition, the current study proposes an FL-based facial emotion recognition system by addressing local imbalance data problems encountered in client devices. First, the local imbalance problem is mitigated by utilizing the data-level artificial resampling method on the client side. To address the possibility of an adversarial attack using imbalanced data, the local training is equipped with a pre-training check to verify if the data being used is imbalanced above a predefined threshold of imbalance ratio. In case of high imbalance, a pre-training step will balance the data locally without sharing any information with other participants thereby ensuring privacy in the FL framework. Experiments have been conducted by using benchmark facial emotion recognition data with a balanced testing strategy. It indicated that considerable improvement can be achieved by the proposed FL-based facial emotion recognition model.
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 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".