@article{ku-etal-2026-structure,
    title = "When Structure Matters: Cross-Lingual Hyperbolic Embeddings for {C}hinese and {E}nglish Wordnets",
    author = "Ku, Mao-Chang  and
      Lian, Da-Chen  and
      Chen, Pin-Er  and
      Wang, Po-Ya Angela  and
      Chen, Wei-Ling  and
      HSIEH, Shu-Kai",
    editor = "Piperidis, Stelios  and
      Bel, N{\'u}ria  and
      van den Heuvel, Henk  and
      Ide, Nancy  and
      Krek, Simon  and
      Toral, Antonio",
    journal = "International Conference on Language Resources and Evaluation",
    volume = "main",
    month = may,
    year = "2026",
    address = "Palma de Mallorca, Spain",
    publisher = "ELRA Language Resource Association",
    url = "https://preview.aclanthology.org/ingest-lrec/2026.lrec-main.944/",
    pages = "12054--12071",
    abstract = "Hyperbolic embeddings such as the Poincar{\'e} model effectively represent lexical hierarchies with low distortion, yet their cross-lingual generalizability remains largely unexplored. This study investigates cross-lingual transfer by training 20-dimensional Poincar{\'e} embeddings exclusively on Open English WordNet (OEWN) hypernymy relations and evaluating on aligned Chinese Wordnet (CWN) synsets under a vocabulary-constrained transfer setting, where CWN-relevant synsets appear in OEWN training data but no Chinese-language supervision is used. We report robust statistical evidence based on the final 10 training checkpoints: Poincar{\'e} embeddings achieve 2.57{\texttimes} higher Mean Reciprocal Rank (MRR) than Euclidean embeddings on CWN (0.030 {\ensuremath{\pm}} 0.001 vs 0.012 {\ensuremath{\pm}} 0.000, p {\ensuremath{<}} 0.001, Cohen{'}s d = 34.48) and 5.61{\texttimes} higher on OEWN (0.016 {\ensuremath{\pm}} 0.000 vs 0.003 {\ensuremath{\pm}} 0.000, p {\ensuremath{<}} 0.001, d = 42.48). Furthermore, hierarchical filtering leveraging the radial dimension of hyperbolic space provides substantial additional gains: +74.6{\%} MRR improvement on CWN and +25.8{\%} on OEWN (both p {\ensuremath{<}} 0.001). The model achieves higher absolute performance on the zero-shot CWN test set (MRR = 0.052 {\ensuremath{\pm}} 0.002) than on the in-domain OEWN test set (MRR = 0.020 {\ensuremath{\pm}} 0.001). We attribute this to structural alignment: CWN{'}s broader branching factor (4.32 vs 1.10) and moderate depth naturally suit hyperbolic geometry{'}s capacity to compactly represent hierarchies. Our findings demonstrate that geometric properties learned from English hypernymy transfer robustly across languages when semantic structures align. We release the aligned CWN{--}OEWN hypernymy evaluation dataset and complete evaluation framework to facilitate future research on geometry-based cross-lingual semantic modeling."
}