Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2023
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2309.16575 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929238521675776 |
|---|---|
| author | Tanzer, Garrett Suzgun, Mirac Visser, Eline Jurafsky, Dan Melas-Kyriazi, Luke |
| author_facet | Tanzer, Garrett Suzgun, Mirac Visser, Eline Jurafsky, Dan Melas-Kyriazi, Luke |
| contents | Large language models (LLMs) can perform impressive feats with in-context learning or lightweight finetuning. It is natural to wonder how well these models adapt to genuinely new tasks, but how does one find tasks that are unseen in internet-scale training sets? We turn to a field that is explicitly motivated and bottlenecked by a scarcity of web data: low-resource languages. In this paper, we introduce MTOB (Machine Translation from One Book), a benchmark for learning to translate between English and Kalamang -- a language with less than 200 speakers and therefore virtually no presence on the web -- using several hundred pages of field linguistics reference materials. This task framing is novel in that it asks a model to learn a language from a single human-readable book of grammar explanations, rather than a large mined corpus of in-domain data, more akin to L2 learning than L1 acquisition. We demonstrate that baselines using current LLMs are promising but fall short of human performance, achieving 44.7 chrF on Kalamang to English translation and 45.8 chrF on English to Kalamang translation, compared to 51.6 and 57.0 chrF by a human who learned Kalamang from the same reference materials. We hope that MTOB will help measure LLM capabilities along a new dimension, and that the methods developed to solve it could help expand access to language technology for underserved communities by leveraging qualitatively different kinds of data than traditional machine translation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2309_16575 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | A Benchmark for Learning to Translate a New Language from One Grammar Book Tanzer, Garrett Suzgun, Mirac Visser, Eline Jurafsky, Dan Melas-Kyriazi, Luke Computation and Language Large language models (LLMs) can perform impressive feats with in-context learning or lightweight finetuning. It is natural to wonder how well these models adapt to genuinely new tasks, but how does one find tasks that are unseen in internet-scale training sets? We turn to a field that is explicitly motivated and bottlenecked by a scarcity of web data: low-resource languages. In this paper, we introduce MTOB (Machine Translation from One Book), a benchmark for learning to translate between English and Kalamang -- a language with less than 200 speakers and therefore virtually no presence on the web -- using several hundred pages of field linguistics reference materials. This task framing is novel in that it asks a model to learn a language from a single human-readable book of grammar explanations, rather than a large mined corpus of in-domain data, more akin to L2 learning than L1 acquisition. We demonstrate that baselines using current LLMs are promising but fall short of human performance, achieving 44.7 chrF on Kalamang to English translation and 45.8 chrF on English to Kalamang translation, compared to 51.6 and 57.0 chrF by a human who learned Kalamang from the same reference materials. We hope that MTOB will help measure LLM capabilities along a new dimension, and that the methods developed to solve it could help expand access to language technology for underserved communities by leveraging qualitatively different kinds of data than traditional machine translation. |
| title | A Benchmark for Learning to Translate a New Language from One Grammar Book |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2309.16575 |