Abstract
The SIGTYP 2022 shared task concerns the problem of word reflex generation in a target language, given cognate words from a subset of related languages. We present two systems to tackle this problem, covering two very different modeling approaches. The first model extends transformer-based encoder-decoder sequence-to-sequence modeling, by encoding all available input cognates in parallel, and having the decoder attend to the resulting joint representation during inference. The second approach takes inspiration from the field of image restoration, where models are tasked with recovering pixels in an image that have been masked out. For reflex generation, the missing reflexes are treated as “masked pixels” in an “image” which is a representation of an entire cognate set across a language family. As in the image restoration case, cognate restoration is performed with a convolutional network.- Anthology ID:
- 2022.sigtyp-1.9
- Volume:
- Proceedings of the 4th Workshop on Research in Computational Linguistic Typology and Multilingual NLP
- Month:
- July
- Year:
- 2022
- Address:
- Seattle, Washington
- Editors:
- Ekaterina Vylomova, Edoardo Ponti, Ryan Cotterell
- Venue:
- SIGTYP
- SIG:
- SIGTYP
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 70–79
- Language:
- URL:
- https://aclanthology.org/2022.sigtyp-1.9
- DOI:
- 10.18653/v1/2022.sigtyp-1.9
- Cite (ACL):
- Christo Kirov, Richard Sproat, and Alexander Gutkin. 2022. Mockingbird at the SIGTYP 2022 Shared Task: Two Types of Models for the Prediction of Cognate Reflexes. In Proceedings of the 4th Workshop on Research in Computational Linguistic Typology and Multilingual NLP, pages 70–79, Seattle, Washington. Association for Computational Linguistics.
- Cite (Informal):
- Mockingbird at the SIGTYP 2022 Shared Task: Two Types of Models for the Prediction of Cognate Reflexes (Kirov et al., SIGTYP 2022)
- PDF:
- https://preview.aclanthology.org/landing_page/2022.sigtyp-1.9.pdf