Pith Number

pith:DEJPA3Z5

pith:2022:DEJPA3Z57HRBELNYKPSFPHSMQD

not attested not anchored not stored refs resolved

No Language Left Behind: Scaling Human-Centered Machine Translation

Alexandre Mourachko, Al Youngblood, Angela Fan, Anna Sun, Bapi Akula, Chau Tran, Christophe Ropers, Cynthia Gao, Daniel Licht, Dirk Rowe, Elahe Kalbassi, Francisco Guzm\'an, Gabriel Mejia Gonzalez, Guillaume Wenzek, Holger Schwenk, James Cross, Janice Lam, Jean Maillard, Jeff Wang (NLLB Team), John Hoffman, Kaushik Ram Sadagopan, Kenneth Heafield, Kevin Heffernan, Loic Barrault, Maha Elbayad, Marta R. Costa-juss\`a, Necip Fazil Ayan, NLLB Team, Onur \c{C}elebi, Philipp Koehn, Pierre Andrews, Prangthip Hansanti, Safiyyah Saleem, Semarley Jarrett, Sergey Edunov, Shannon Spruit, Shruti Bhosale, Skyler Wang, Vedanuj Goswami

A sparsely gated mixture of experts model trained on mined low-resource data achieves 44% relative BLEU improvement in translating 200 languages.

arxiv:2207.04672 v3 · 2022-07-11 · cs.CL · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{DEJPA3Z57HRBELNYKPSFPHSMQD}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Our model achieves an improvement of 44% BLEU relative to the previous state-of-the-art, laying important groundwork towards realizing a universal translation system.

C2weakest assumption

The novel data mining techniques and architectural/training improvements produce genuinely higher-quality and safer translations for low-resource languages rather than merely fitting the new benchmark or human raters.

C3one line summary

A sparsely gated mixture-of-experts model trained on newly mined low-resource data achieves 44% relative BLEU improvement across 200 languages while adding human safety evaluation.

References

16 extracted · 16 resolved · 2 Pith anchors

[1] URL https://arxiv.org/abs/2110.03036. Benjamin Akera, Jonathan Mukiibi, Lydia Sanyu Naggayi, Claire Babirye, Isaac Owomugisha, Solomon Nsumba, Joyce Nakatumba-Nabende, Engineer Bainomugisha, Ernest Mw

[2] Farhad Akhbardeh, Arkady Arkhangorodsky, Magdalena Biesialska, Ondřej Bojar, Rajen Chatterjee, Vishrav Chaudhary, Marta R 2021

[3] In: Zong, C., Xia, F., Li, W., Navigli, R 2016 · doi:10.18653/v1/

[4] doi: 10.18653/v1/2021.iwslt-1.1 2021 · doi:10.18653/v1/2021.iwslt-1.1

[5] Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation 2022 · doi:10.18653/v1/2020.acl-main.485

Formal links

2 machine-checked theorem links

Cited by

100 papers in Pith

Audio Sentiment Analysis via Distillation and Cross-Modal Integration of Generated Multilingual Transcripts

Andha-Dhun: A First Look at Audio Descriptions in Hindi

The Tatoxa System for Text Detoxification in Low-Resource Languages: The Case of Tatar

SARA: Unlocking Multilingual Knowledge in Mixture-of-Experts via Semantically Anchored Routing Alignment

Multilingual Hematology Visual Question Answering Dataset

Receipt and verification

First computed	2026-07-05T04:51:35.136376Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

1912f06f3df9e2122db853e4579e4c80e77f0c8c251724fb6bec603e3d111512

Aliases

arxiv: 2207.04672 · arxiv_version: 2207.04672v3 · doi: 10.48550/arxiv.2207.04672 · pith_short_12: DEJPA3Z57HRB · pith_short_16: DEJPA3Z57HRBELNY · pith_short_8: DEJPA3Z5

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/DEJPA3Z57HRBELNYKPSFPHSMQD \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 1912f06f3df9e2122db853e4579e4c80e77f0c8c251724fb6bec603e3d111512

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "f38df21833913da5e913b5c418119c400fbbc9f7ebbf02bbc3229f86d01400be",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by-sa/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2022-07-11T07:33:36Z",
    "title_canon_sha256": "01ed8c3f5073036312b0a2e1fc34f8dfd220e1d0182d756cd9061ffa8c81e502"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2207.04672",
    "kind": "arxiv",
    "version": 3
  }
}