Pith Number

pith:R5MPPVO6

pith:2026:R5MPPVO6UYQITCVRWN65DUXVSS

not attested not anchored not stored refs resolved

WARDEN: Endangered Indigenous Language Transcription and Translation with 6 Hours of Training Data

Liang Zheng, Naijing Liu, Yunzhong Hou, Ziheng Zhang

A two-stage pipeline transcribes and translates the endangered Wardaman language using only six hours of annotated data.

arxiv:2605.13846 v1 · 2026-05-13 · 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{R5MPPVO6UYQITCVRWN65DUXVSS}

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

Using a mere 6 hours of annotated data, WARDEN outperforms larger open-source and proprietary models and establishes a strong baseline.

C2weakest assumption

That the two-stage pipeline with Sundanese initialization and dictionary-guided LLM reasoning will reliably outperform unified models in extremely low-data regimes without overfitting or domain mismatch.

C3one line summary

WARDEN achieves better transcription and translation for Wardaman than larger models by separating the tasks and using Sundanese initialization plus a domain dictionary with just 6 hours of data.

References

23 extracted · 23 resolved · 2 Pith anchors

[1] WARDEN: Endangered Indigenous Language Transcription and Translation with 6 Hours of Training Data 2025 · arXiv:2605.13846

[2] Related Work Recent studies have shown that fine-tuning models like Whis- per for translation and transcription require a significant amount of data even when dealing with low-resource languages. Liu

[3] 1, the proposed W ARDEN system is com- posed of two separate stages: a transcription stage and a trans- lation stage

[4] Transcription:{transcription}. Lexicon en- tries:{lexicon entries}

[5] Dataset Data in this paper comes from a long-term anthropological lin- guistic documentation project on the Wardaman language from 1976 to 2025 1976

Cited by

1 paper in Pith

WARDEN: Endangered Indigenous Language Transcription and Translation with 6 Hours of Training Data

Receipt and verification

First computed	2026-05-18T02:44:09.347264Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

8f58f7d5dea620898ab1b37dd1d2f59497c7c6ddbad36c604dfd4e6613f93dc2

Aliases

arxiv: 2605.13846 · arxiv_version: 2605.13846v1 · doi: 10.48550/arxiv.2605.13846 · pith_short_12: R5MPPVO6UYQI · pith_short_16: R5MPPVO6UYQITCVR · pith_short_8: R5MPPVO6

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/R5MPPVO6UYQITCVRWN65DUXVSS \
  | 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: 8f58f7d5dea620898ab1b37dd1d2f59497c7c6ddbad36c604dfd4e6613f93dc2

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "62acc9da46ef10fd8f8337152667d23794d8d3aaaf85bd54abd2ab4057369a2d",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-13T17:59:52Z",
    "title_canon_sha256": "fabe073c790c2c92566cbfcb7de748b9fa4159520877aedf897b618b261c0727"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13846",
    "kind": "arxiv",
    "version": 1
  }
}