Investigating Biases in COVID-19 Diagnostic Systems Processed with Automated Speech Anonymization Algorithms
Bibliographic record
Abstract
Automated voice anonymization algorithms are used to obfuscate speaker identity while leaving other vocal attributes untouched; they have been used for e.g., speech recognition, speech emotion detection, and most recently, remote speechbased health diagnostics.However, speech data is commonly collected in an uncontrolled manner in various environments, potentially compromising its quality, and frequently omits key metadata that could improve model performance.In this study, we employed the Cambridge COVID-19 sound database and used COVID-19 detection as a case study.We first present descriptive statistics on sample composition (i.e., COVID-19 status, age, gender).We also present a measure of signal-tonoise ratio (SNR), a feature of speech that can denote individuals' socioeconomic status.Next, we assess how age and SNR, the two most unbalanced features of the dataset, are associated with model performance and the impact of automated anonymization algorithms performance.Our findings suggest the existence of diagnostic biases related to age and SNR of the recording, which become more prominent after anonymization.To tackle these biases, we explore the usefulness of two data augmentation methods.We show that although data augmentation helps to recover some loss in overall performance, it can lead to a larger discrepancy in performance for over-represented and under-represented groups.We conclude with a discussion of the limitations associated with using SNR as an indicator of socioeconomic status, and of the potential effects of diagnostic biases associated with socioeconomic status.
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 | 0.000 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 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) | 0.000 | 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".