# Cross-Lingual Syntactic Transfer through Unsupervised Adaptation of Invertible Projections

arXiv: https://arxiv.org/abs/1906.02656 · HTML: https://mlsanity.com/paper/1906.02656/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Jun 2019.

## Built on by (most cited first, 7 of 7)

- [PPT: Parsimonious Parser Transfer for Unsupervised Cross-Lingual Adaptation](https://mlsanity.com/paper/2101.11216/) (2021)
- [Substructure Distribution Projection for Zero-Shot Cross-Lingual Dependency Parsing](https://mlsanity.com/paper/2110.08538/) (2021)
- [Boosting Zero-shot Cross-lingual Retrieval by Training on Artificially Code-Switched Data](https://mlsanity.com/paper/2305.05295/) (2023)
- [To token or not to token: A Comparative Study of Text Representations for Cross-Lingual Transfer](https://mlsanity.com/paper/2310.08078/) (2023)
- [Unsupervised Cross-Lingual Transfer of Structured Predictors without Source Data](https://mlsanity.com/paper/2110.03866/) (2021)
- [Improving Low-Resource Cross-lingual Parsing with Expected Statistic Regularization](https://mlsanity.com/paper/2210.09428/) (2022)
- [Information Loss in LLMs' Multilingual Translation: The Role of Training Data, Language Proximity, and Language Family](https://mlsanity.com/paper/2506.23340/) (2025)

## Abstract (arXiv)

Cross-lingual transfer is an effective way to build syntactic analysis tools in low-resource languages. However, transfer is difficult when transferring to typologically distant languages, especially when neither annotated target data nor parallel corpora are available. In this paper, we focus on methods for cross-lingual transfer to distant languages and propose to learn a generative model with a structured prior that utilizes labeled source data and unlabeled target data jointly. The parameters of source model and target model are softly shared through a regularized log likelihood objective. An invertible projection is employed to learn a new interlingual latent embedding space that compensates for imperfect cross-lingual word embedding input. We evaluate our method on two syntactic tasks: part-of-speech (POS) tagging and dependency parsing. On the Universal Dependency Treebanks, we use English as the only source corpus and transfer to a wide range of target languages. On the 10 languages in this dataset that are distant from English, our method yields an average of 5.2% absolute improvement on POS tagging and 8.3% absolute improvement on dependency parsing over a direct transfer method using state-of-the-art discriminative models.