Learning the Arabic Dialect Continuum as a Continuous Space: A Regression Approach to Speaker Origin Prediction
We present a regression-based approach to Arabic dialect geolocation that models dialectal variation as a continuous geographic space rather than discrete categories. Speaker origin is predicted as continuous latitude-longitude coordinates using a hierarchical neural architecture that fuses frame-level XLS-R-300M and Whisper-large-v3 encoder representations with phonotactic descriptors through a Transformer encoder and a learnable attention-pooled query. A spherical geodesic loss directly optimi
Researchers present a regression-based approach to Arabic dialect geolocation, predicting speaker origin as continuous latitude-longitude coordinates with a median localization error of 481.2 km. The model uses a hierarchical neural architecture and achieves 64.5% and 45.2% accuracy for country and city predictions, respectively.
Summarised by netranta from News. Open the original for the full story.
Observations (0)
Log in to add an observation.
No observations yet — add the first.