Herygers, Aaricia (2022) Spraakherkenning, wa is da? — Bias in Flemish Speech Recognition. Master thesis, Voice Technology (VT).
|
PDF
MA 4951638 A Herygers.pdf Download (1MB) | Preview |
Abstract
Sociolinguistic factors such as age (Vipperla, Renals, & Frankel, 2008) and gender (Tatman, 2017) have been shown to impact the performance of various automatic speech recognition (ASR) models. Previous research has touched upon such performance discrepancies, uncovering biases in ASR models, but has often focused on the English language (e.g., Kathania, Reddy Kadiri, Alku, & Kurimo, 2020; Tatman & Kasten, 2017; Vipperla, Renals, & Frankel, 2010). However, as these systems are used worldwide, finding biases in different languages is of high importance. With this thesis, I extend recent research by Feng, Kudina, Halpern, and Scharenborg (2021), who sought to find biases based on age, region, gender, and non-nativeness in a Dutch ASR model. Like Feng et al. (2021), I use the Netherlandic Dutch data from the Spoken Dutch Corpus (Oostdijk, 2000) to train a hybrid deep neural network-hidden Markov model (DNN-HMM). However, the previous study did not take into account the various regional variants of Belgian Dutch, which is also known as Flemish, e.g., West Flemish, and Brabantian (Odijk, 2012). I therefore evaluate the model using the Flemish data from the JASMIN-CGN corpus (Cucchiarini, Van hamme, van Herwijnen, & Smits, 2006). The evaluation confirms a bias against speakers from West Flanders and Limburg, as well as against children, male speakers, and non-native speakers. In addition, the discussion of the findings includes an analysis of the most misrecognized phonemes. The current study contributes to a better understanding of bias, and subsequently inclusivity, in ASR.
| Item Type: | Thesis (Master) |
|---|---|
| Name supervisor: | Verkhodanova, V. and Coler, M.L. |
| Date Deposited: | 09 Sep 2022 09:00 |
| Last Modified: | 09 Sep 2022 09:00 |
| URI: | https://campus-fryslan.studenttheses.ub.rug.nl/id/eprint/231 |
Actions (login required)
 |
View Item |