ACES: Accent Subspaces for Coupling, Explanations, and Stress-Testing in Automatic Speech Recognition
ASR systems exhibit persistent performance disparities across accents, but whether these gaps reflect superficial biases or deep structural vulnerabilities remains unclear. We introduce ACES, a three-stage audit that extracts accent-discriminative subspaces from ASR representations, constrains adversarial attacks to them, and tests whether removing them improves fairness. On Wav2Vec2-base with seven accents, imperceptible perturbations (~60 dB SNR) along the accent subspace amplify the WER disparity gap by nearly 50% (21.3->31.8 pp), exceeding random-subspace controls; a permuted-label test confirms specificity to genuine accent structure. Partially removing the subspace worsens both WER and disparity, revealing that accent-discriminative and recognition-critical features are deeply entangled. ACES thus positions accent subspaces as powerful fairness-auditing tools, not simple erasure levers.
Code (0)
등록된 구현이 없습니다.
Tasks
Speech RecognitionSimilar Papers 제목 키워드 기반
Explaining Aggregates for Exploratory Analytics
Analysts wishing to explore multivariate data spaces, typically pose queries involving selection operators, i.e., range or radius queries, which define data subspaces of possible interest and then use aggregation functio…
Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspaces
Explainable AI aims to overcome the black-box nature of complex ML models like neural networks by generating explanations for their predictions. Explanations often take the form of a heatmap identifying input features (e…
Emirati-Accented Speaker Identification in Stressful Talking Conditions
This research is dedicated to improving text-independent Emirati-accented speaker identification performance in stressful talking conditions using three distinct classifiers: First-Order Hidden Markov Models (HMM1s), Sec…
Speaker IdentificationAD-MERCS: Modeling Normality and Abnormality in Unsupervised Anomaly Detection
Most anomaly detection systems try to model normal behavior and assume anomalies deviate from it in diverse manners. However, there may be patterns in the anomalies as well. Ideally, an anomaly detection system can explo…
Anomaly DetectionUnsupervised Anomaly DetectionTri-Subspaces Disentanglement for Multimodal Sentiment Analysis
Multimodal Sentiment Analysis (MSA) integrates language, visual, and acoustic modalities to infer human sentiment. Most existing methods either focus on globally shared representations or modality-specific features, whil…
Multimodal Intent RecognitionMultimodal Sentiment Analysis