TURNA: A Turkish Encoder-Decoder Language Model for Enhanced Understanding and Generation
Findings of the 62nd Annual Meeting of the Association for Computational Linguistics, ACL 2024, Hybrid, Bangkok, Tayland, 11 - 16 Ağustos 2024, ss.10103-10117, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.18653/v1/2024.findings-acl.600
- Basıldığı Şehir: Hybrid, Bangkok
- Basıldığı Ülke: Tayland
- Sayfa Sayıları: ss.10103-10117
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Boğaziçi Üniversitesi Adresli: Evet
Özet
The recent advances in natural language processing have predominantly favored well-resourced English-centric models, resulting in a significant gap with low-resource languages. In this work, we introduce TURNA, a language model developed for the low-resource language Turkish and is capable of both natural language understanding and generation tasks. TURNA is pretrained with an encoder-decoder architecture based on the unified framework UL2 with a diverse corpus that we specifically curated for this purpose. We evaluated TURNA with three generation and five understanding tasks for Turkish. The results show that TURNA outperforms several multilingual models in both understanding and generation tasks, and competes with monolingual Turkish models in understanding tasks.