Linear Semantic Segmentation for Low-Resource Spoken DialectsKirill ChirkunovYounes Samihet al.2026ACL 2026Paper