Papers › Neural Machine Translation of Rare Words with Subword Units

Neural Machine Translation of Rare Words with Subword Units

31 Aug 2015ACL 2016 8arXiv:1508.07909archive 2025-07-28

Rico Sennrich, Barry Haddow, Alexandra Birch

Neural machine translation (NMT) models typically operate with a fixed vocabulary, but translation is an open-vocabulary problem. Previous work addresses the translation of out-of-vocabulary words by backing off to a dictionary. In this paper, we introduce a simpler and more effective approach, making the NMT model capable of open-vocabulary translation by encoding rare and unknown words as sequences of subword units. This is based on the intuition that various word classes are translatable via smaller units than words, for instance names (via character copying or transliteration), compounds (via compositional translation), and cognates and loanwords (via phonological and morphological transformations). We discuss the suitability of different word segmentation techniques, including simple character n-gram models and a segmentation based on the byte pair encoding compression algorithm, and empirically show that subword models improve over a back-off dictionary baseline for the WMT 15 translation tasks English-German and English-Russian by 1.1 and 1.3 BLEU, respectively.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="1508.07909")

Code

Syntology Ran 21 of 30 code samples harvested from 7 repositories linked to this paper; 9 have no recorded run. Of those that ran: 17 ran · our draft was wrong; 1 ran · fixture could not drive it; 3 ran with no contract checked.

By repository: official repository: 4 samples from 1 repository, 2 ran; community (archive-listed): 24 samples from 6 repositories, 17 ran; 2 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

26 repositories listed; official and paper-mentioned ones first.

Automattic/wp-translate mentioned on GitHubtf report
Avmb/code-docstring-corpus mentioned on GitHub report
ThAIKeras/bert mentioned on GitHubtf report
Xinsen-Zhang/transformer mentioned on GitHubpaddle report
facebookresearch/fairseq mentioned on GitHubpytorchMIT report
glample/fastBPE mentioned on GitHubMIT report
johnr0/TaleBrush-backend mentioned on GitHubpytorch report
karpathy/minbpe mentioned on GitHub report
kh-mo/QA_wikisql mentioned on GitHub report
lkfo415579/MT-Readling-List mentioned on GitHubtf report
nyu-dl/dl4mt-c2c mentioned on GitHubBSD-3-Clause report
nyu-dl/dl4mt-cdec mentioned on GitHubBSD-3-Clause report
nyu-dl/dl4mt-simul-trans mentioned on GitHub report
salesforce/GeDi mentioned on GitHubpytorchBSD-3-Clause report
simonjisu/NMT mentioned on GitHubpytorchMIT report
siyuofzhou/CNNSeqToSeq mentioned on GitHubtorch report
thinkwee/DPP_CNN_Summarization mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

30 samples harvested; 21 ran; 0 honoured the contract we drafted; 9 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

17ran · our draft was wrong
1ran · fixture could not drive it
3ran
9unverified

Licence: 21 of the 30 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 7 repositories linked to this paper, official or community; each sample names its own and says which. Some samples are identical code Syntology first harvested from another repository; for those, this paper's copy is not located and its licence is not recorded. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

open_file rsennrich/subword-nmt/subword_nmt/learn_bpe.py official repository ran · our draft was wrong MIT (permissive) · 6aafe96a739d0b09 · report
replace_pair rsennrich/subword-nmt/subword_nmt/learn_bpe.py official repository ran MIT (permissive) · ab293eff6ffc4b69 · report
_get_vocabulary rsennrich/subword-nmt/subword_nmt/learn_bpe.py official repository unverified MIT (permissive) · 800b6a080cb3c624 · report
learn_bpe rsennrich/subword-nmt/subword_nmt/learn_bpe.py official repository unverified MIT (permissive) · d86dc2151a1bd58e · report
BPE ThAIKeras/bert/bpe_helper.py community (archive-listed) ran fingerprinted Apache-2.0 (permissive) · 33b1567b39f86e71 · report
_bpe2words nyu-dl/dl4mt-simul-trans/simultrans_model.py community (archive-listed) ran · our draft was wrong BSD-3-Clause (permissive) · 70a8fed9225c5845 · report
check_merge_info SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 2032a0c86da75680 · report
delete_some_stats SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · fixture could not drive it no licence file found · pointer only · c837ac919bdfe91e · report
get_pair_statistics kh-mo/QA_wikisql/learn_bpe.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · e27a7ada9b96b37e · report
get_stats SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · afdbbe2995c004e2 · report
get_vocabulary kh-mo/QA_wikisql/learn_bpe.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 9425aef3e44f6979 · report
get_vocabulary thinkwee/DPP_CNN_Summarization/fairseq/raw_dataset/subword_nmt/learn_bpe.py community (archive-listed) ran no licence file found · pointer only · bc7574327ea1f634 · report
get_vocabulary_from_learn_BPE SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · afac3ba7a001359e · report
learn_bpe kh-mo/QA_wikisql/learn_bpe.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · a4985508aef2cb2a · report
merge_bpe_word SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 029694838a73e5fc · report
merge_dictionary SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 72ccea7ec0eb1e6a · report
render_token karpathy/minbpe/minbpe/base.py community (archive-listed) ran · our draft was wrong MIT (permissive) · 1c202c7e568439b5 · report
replace_control_characters karpathy/minbpe/minbpe/base.py community (archive-listed) ran · our draft was wrong fingerprinted MIT (permissive) · 15d9c32bbe7da76f · report
replace_pair kh-mo/QA_wikisql/learn_bpe.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · d1027b3c5283677a · report
replace_pair thinkwee/DPP_CNN_Summarization/fairseq/raw_dataset/subword_nmt/learn_bpe.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · 684c47713cee0e44 · report
selective_get_stats SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · c0d8d4efee4a4015 · report
Tokenizer karpathy/minbpe/minbpe/base.py community (archive-listed) unverified MIT (permissive) · 293d3695d53b0cb1 · report
_learn_bpe SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) unverified no licence file found · pointer only · 133f6dc75e90bbd6 · report
learn_bpe thinkwee/DPP_CNN_Summarization/fairseq/raw_dataset/subword_nmt/learn_bpe.py community (archive-listed) unverified no licence file found · pointer only · 0c8c96b6d6bd911a · report
prune_stats thinkwee/DPP_CNN_Summarization/fairseq/raw_dataset/subword_nmt/learn_bpe.py community (archive-listed) unverified no licence file found · pointer only · 927c3c1d2631e415 · report
save_voca SeonbeomKim/Python-Byte_Pair_Encoding/bpe_module/learn_BPE.py community (archive-listed) unverified no licence file found · pointer only · e2122ad711c160ec · report
update_pair_statistics kh-mo/QA_wikisql/learn_bpe.py community (archive-listed) unverified no licence file found · pointer only · c5fcb9cba4915130 · report
update_pair_statistics thinkwee/DPP_CNN_Summarization/fairseq/raw_dataset/subword_nmt/learn_bpe.py community (archive-listed) unverified no licence file found · pointer only · 718aeb6813392ac3 · report
get_pair_statistics identical code first harvested elsewhere ran · our draft was wrong licence of this copy not recorded · fe96394f96a9c638 · report
get_vocabulary identical code first harvested elsewhere ran · our draft was wrong licence of this copy not recorded · b83d785ca99bfc9f · report

Tasks

Machine TranslationNMTSegmentationTranslation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Machine Translation WMT2015 English-German BPE word segmentation BLEU score 22.8 #4 of 6 Archive leaderboard report
Machine Translation WMT2015 English-Russian C2-50k Segmentation BLEU score 20.9 #1 of 1 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: BPE

BPE

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections