arXiv cs.CL· Yuliang Liu, Xinyu Wang, Xiang Bai·· 11 小时前AI 评分38
研究发现：让未见语言进入共享语义空间的是熟悉的字母而非亲属语言
Familiar letters, not related languages, let unseen languages into a shared space of meaning
AI 导读
无分词器的视觉 Transformer GlyphMind 将文本作为像素读取，在 27 种书写系统上预训练对齐译文后，仅用未翻译的老挝语适配，检索英文译文的频率达到最强简单基线的七倍。
正文 · 原文
Abstract:Hundreds of languages have written text but almost no translations, and whether a model learning one from untranslated text builds on a related language or familiar letters is unclear: they usually coincide. Here we show that familiar letters, not a related language, let an unseen language into a shared space of meaning. GlyphMind, a tokenizer-free vision transformer that reads text as pixels and lets any sentence be shown in any script, was pretrained to align translations across 27 writing systems. Adapted on untranslated Lao, a language and script it had never seen, it retrieved English translations seven times as often as the strongest simple baseline. Rewritten in unfamiliar letters, Lao did not enter although its close relative Thai was in pretraining; Santali, which had none, entered in Devanagari, familiar from Hindi, but not in its own script. Transliteration into familiar scripts may thus open multilingual models to languages without translations.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.06006 [cs.CL] |
| (or arXiv:2611.06006v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06006 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuliang Liu [view email]
[v1]
Wed, 4 Nov 2026 05:54:23 UTC (2,207 KB)
来源：arXiv cs.CL · arxiv.org