analiza aiqbench
This research introduces Symphonym, a model that unifies toponym matching across 36 writing systems by learning from articulatory features rather than relying on language-specific algorithms or romanization. By distilling knowledge from IPA transcriptions into a character-level student network, the system achieves high recall on medieval Hebrew and Arabic datasets while correcting previous biases in handling Chinese characters with Japanese readings. The study provides critical insights into the trade-offs between cross-script generalization and string-metric performance within Latin scripts, offering a robust solution for integrating diverse multilingual geographic sources.
Symphonym: Universal Phonetic Embeddings for Cross-Script Toponym Matching
arXiv:2601.06932v5 Announce Type: replace-cross Abstract: Matching place names across writing systems is a persistent obstacle to integrating multilingual geographic sources, from modern gazetteers to medieval itineraries and colonial-era surveys. Existing approaches rely on language-specific phonetic algorithms or on romanisation that discards phonetic information, and none generalises across scripts. Symphonym maps toponyms from thirty-six writing systems into a unified 128-dimensional phonetic space, enabling direct cross-script comparison without language identification or phonetic resources at inference time. A Teacher-Student distillation architecture learns from articulatory features of IPA transcriptions and transfers this knowledge to a character-level Student. Trained on 73.5 million toponyms from GeoNames, Wikidata and the Getty TGN, the Student achieves the highest Recall@1 (89.3%) and MRR (92.8%) on the MEHDIE benchmark of medieval Hebrew and Arabic toponym matches, which is independent of the training data. An ablation on raw articulatory features alone reaches only 45.0% MRR. This revision reports a second model generation and three results that qualify the first: a corpus defect caused the original system to learn Chinese characters with Japanese readings, which we correct and quantify; the encoder weighted character content far above order, admitting 70.5% of random anagrams past the retrieval gate, which targeted negatives reduce to 5.0% without loss of typo tolerance; and better retrieval came with slightly worse separation of true from false matches. We report where the method wins decisively (across scripts) and where string metrics remain preferable (within the Latin script), and document an independent out-of-domain deployment on archival personal names.
Abstract:Matching place names across writing systems is a persistent obstacle to integrating multilingual geographic sources, from modern gazetteers to medieval itineraries and colonial-era surveys. Existing approaches rely on language-specific phonetic algorithms or on romanisation that discards phonetic information, and none generalises across scripts. Symphonym maps toponyms from thirty-six writing systems into a unified 128-dimensional phonetic space, enabling direct cross-script comparison without language identification or phonetic resources at inference time. A Teacher-Student distillation architecture learns from articulatory features of IPA transcriptions and transfers this knowledge to a character-level Student. Trained on 73.5 million toponyms from GeoNames, Wikidata and the Getty TGN, the Student achieves the highest Recall@1 (89.3%) and MRR (92.8%) on the MEHDIE benchmark of medieval Hebrew and Arabic toponym matches, which is independent of the training data. An ablation on raw articulatory features alone reaches only 45.0% MRR. This revision reports a second model generation and three results that qualify the first: a corpus defect caused the original system to learn Chinese characters with Japanese readings, which we correct and quantify; the encoder weighted character content far above order, admitting 70.5% of random anagrams past the retrieval gate, which targeted negatives reduce to 5.0% without loss of typo tolerance; and better retrieval came with slightly worse separation of true from false matches. We report where the method wins decisively (across scripts) and where string metrics remain preferable (within the Latin script), and document an independent out-of-domain deployment on archival personal names.
25 pages, 1 figure, 6 tables. v5: revised for the second model generation (Symphonym v8); corrects the CJK-Hiragana explanation given in v1-v4. Models and data: this https URL
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
I.2.7; I.5.1; H.2.8
arXiv:2601.06932 [cs.CL]
(or arXiv:2601.06932v5 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2601.06932 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 369 cuvinte
Subiecte conexe
Introducing the Ettin Reranker Family
Hugging Face · 19 mai 2026
[AINews] The Future of Latent Space
Latent Space · 25 sept. 2026
Multi-Vector (Late Interaction) Embedding Models with Sentence Transformers
Hugging Face · 18 aug. 2026
Beyond Scale and Generation: Understanding Language Model-based Entity Matching
arXiv · 28 iul. 2026