We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Using Perturbed Length-aware Positional Encoding for Non-autoregressive Neural Machine Translation

Abstract: Non-autoregressive neural machine translation (NAT) usually employs sequence-level knowledge distillation using autoregressive neural machine translation (AT) as its teacher model. However, a NAT model often outputs shorter sentences than an AT model. In this work, we propose sequence-level knowledge distillation (SKD) using perturbed length-aware positional encoding and apply it to a student model, the Levenshtein Transformer. Our method outperformed a standard Levenshtein Transformer by 2.5 points in bilingual evaluation understudy (BLEU) at maximum in a WMT14 German to English translation. The NAT model output longer sentences than the baseline NAT models.
Comments: 5 pages, 1 figures. Will be presented at ACL SRW 2021
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2107.13689 [cs.CL]
  (or arXiv:2107.13689v1 [cs.CL] for this version)

Submission history

From: Yui Oka [view email]
[v1] Thu, 29 Jul 2021 00:51:44 GMT (58kb,D)

Link back to: arXiv, form interface, contact.