Search for: All records

In this paper, we explore automatic prediction of dialect density of the African American English (AAE) dialect, where dialect density is defined as the percentage of words in an utterance that contain characteristics of the non-standard dialect. We investigate several acoustic and language modeling features, including the commonly used X-vector representation and ComParE feature set, in addition to information extracted from ASR transcripts of the audio files and prosodic information. To address issues of limited labeled data, we use a weakly supervised model to project prosodic and X-vector features into low-dimensional task-relevant representations. An XGBoost model is then used to predict the speaker's dialect density from these features and show which are most significant during inference. We evaluate the utility of these features both alone and in combination for the given task. This work, which does not rely on hand-labeled transcripts, is performed on audio segments from the CORAAL database. We show a significant correlation between our predicted and ground truth dialect density measures for AAE speech in this database and propose this work as a tool for explaining and mitigating bias in speech technology. 

Automatic assessment of depression from speech signals is affected by variabilities in acoustic content and speakers. In this study, we focused on addressing these variabilities. We used a database comprised of recordings of interviews from a large number of female speakers: 735 individuals suffering from depressive (dysthymia and major depression) and anxiety disorders (generalized anxiety disorder, panic disorder with or without agoraphobia) and 953 healthy individuals. Leveraging this unique and extensive database, we built an i-vector framework. In order to capture various aspects of speech signals, we used voice quality features in addition to conventional cepstral features. The features (F0, F1, F2, F3, H1-H2, H2-H4, H4-H2k, A1, A2, A3, and CPP) were inspired by a psychoacoustic model of voice quality [1]. An i-vector-based system using Mel Frequency Cepstral Coefficients (MFCCs) and another using voice quality features was developed. Voice quality features performed as well as MFCCs. A score-level fusion was then used to combine these two systems, resulting in a 6% relative improvement in accuracy in comparison with the i-vector system based on MFCCs alone. The system was robust even when the duration of the utterances was shortened to 10 seconds.