Modality Matching Matters: Calibrating Language Distances for Cross-Lingual Transfer in URIEL+
Published in European Chapter of the Association for Computational Linguistics (EACL) Student Research Workshop 2026, 2026
This paper introduces modality-specific language distance representations for cross-lingual transfer, including speaker-weighted geography, hyperbolic genealogy, and latent-variable typology. It combines these signals into a composite distance that improves transfer-language selection across multiple benchmarks.
Designated as a spotlight paper.
Recommended citation: York Hay Ng*, Aditya Khan*, Xiang Lu*, Matteo Salloum, Michael Zhou, Phuong Hanh Hoang, A. Seza Dogruoz, and En-Shiun Annie Lee. 2026. Modality Matching Matters: Calibrating Language Distances for Cross-Lingual Transfer in URIEL+. In Proceedings of the European Chapter of the Association for Computational Linguistics (EACL) Student Research Workshop 2026.
Download Paper