Abstract
We propose several ways of reusing subword embeddings and other weights in subword-aware neural language models. The proposed techniques do not benefit a competitive character-aware model, but some of them improve the performance of syllable- and morpheme-aware models while showing significant reductions in model sizes. We discover a simple hands-on principle: in a multi-layer input embedding model, layers should be tied consecutively bottom-up if reused at output. Our best morpheme-aware model with properly reused weights beats the competitive word-level model by a large margin across multiple languages and has 20%-87% fewer parameters.- Anthology ID:
- N18-1128
- Volume:
- Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers)
- Month:
- June
- Year:
- 2018
- Address:
- New Orleans, Louisiana
- Editors:
- Marilyn Walker, Heng Ji, Amanda Stent
- Venue:
- NAACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 1413–1423
- Language:
- URL:
- https://aclanthology.org/N18-1128
- DOI:
- 10.18653/v1/N18-1128
- Cite (ACL):
- Zhenisbek Assylbekov and Rustem Takhanov. 2018. Reusing Weights in Subword-Aware Neural Language Models. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pages 1413–1423, New Orleans, Louisiana. Association for Computational Linguistics.
- Cite (Informal):
- Reusing Weights in Subword-Aware Neural Language Models (Assylbekov & Takhanov, NAACL 2018)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-2/N18-1128.pdf
- Code
- zh3nis/morph-sum
- Data
- WikiText-2