Abstract
Modular deep learning has been proposed for the efficient adaption of pre-trained models to new tasks, domains and languages. In particular, combining language adapters with task adapters has shown potential where no supervised data exists for a language. In this paper, we explore the role of language adapters in zero-shot cross-lingual transfer for natural language understanding (NLU) benchmarks. We study the effect of including a target-language adapter in detailed ablation studies with two multilingual models and three multilingual datasets. Our results show that the effect of target-language adapters is highly inconsistent across tasks, languages and models. Retaining the source-language adapter instead often leads to an equivalent, and sometimes to a better, performance. Removing the language adapter after training has only a weak negative effect, indicating that the language adapters do not have a strong impact on the predictions.- Anthology ID:
- 2024.moomin-1.4
- Volume:
- Proceedings of the 1st Workshop on Modular and Open Multilingual NLP (MOOMIN 2024)
- Month:
- March
- Year:
- 2024
- Address:
- St Julians, Malta
- Editors:
- Raúl Vázquez, Timothee Mickus, Jörg Tiedemann, Ivan Vulić, Ahmet Üstün
- Venues:
- MOOMIN | WS
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 24–43
- Language:
- URL:
- https://aclanthology.org/2024.moomin-1.4
- DOI:
- Cite (ACL):
- Jenny Kunz and Oskar Holmström. 2024. The Impact of Language Adapters in Cross-Lingual Transfer for NLU. In Proceedings of the 1st Workshop on Modular and Open Multilingual NLP (MOOMIN 2024), pages 24–43, St Julians, Malta. Association for Computational Linguistics.
- Cite (Informal):
- The Impact of Language Adapters in Cross-Lingual Transfer for NLU (Kunz & Holmström, MOOMIN-WS 2024)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-5/2024.moomin-1.4.pdf