Metraipolis › archive › dc_237a246a060f

Synthetic-to-Real Code-Switched ASR: A Spontaneous Nepali-English Podcast Benchmark — arXiv:2610.30860

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.30860
License (as archived)
CC0-1.0
Archived
2026-10-23T15:20:31.493Z
SHA-256 of the archived text
660b5ff4421252dabbd7844acc7637968f506262e0424f43b413265aa99c9ec8
Stable URL of this page
https://metraipolis.com/documento/dc_237a246a060f

How to cite

Synthetic-to-Real Code-Switched ASR: A Spontaneous Nepali-English Podcast Benchmark — arXiv:2610.30860 — arXiv (en), https://arxiv.org/abs/2610.30860 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_237a246a060f (Metraipolis archive dc_237a246a060f, sha256 660b5ff4421252dabbd7844acc7637968f506262e0424f43b413265aa99c9ec8)
BibTeX
@misc{dc237a246a060f,
  title = {Synthetic-to-Real Code-Switched ASR: A Spontaneous Nepali-English Podcast Benchmark — arXiv:2610.30860},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.30860},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_237a246a060f; sha256 660b5ff4421252dabbd7844acc7637968f506262e0424f43b413265aa99c9ec8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30860. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Nepali-English code-switched conversational speech presents severe challenges for automatic speech recognition (ASR), chiefly through script confusion (inappropriate mixing of Devanagari and Latin orthography) and elevated error rates at language switch points. Supervised fine-tuning (SFT) can adapt multilingual foundation models, but human code-switched audio is expensive to collect and annotate. We curate a new benchmark of 13,044 spontaneous podcast utterances (15.45 h total; 10,214 test, 2,830 dev), audited and verified to eliminate machine translation leakages and acoustic loops. We investigate a two-stage synthetic-to-real adaptation framework: Stage 1 applies SFT on 0.95 h of speech synthesized via a multi-speaker, dual-engine TTS pipeline; Stage 2 explores Group Relative Policy Optimization (GRPO) with transcription and script rewards. We find that synthetic SFT provides the primary cross-domain transfer to spontaneous human speech, reducing Mixed Error Rate (MER) from 23.40% to 21.52% (-1.88 pp) and switch-point error rate (SPER) from 35.89% to 28.77% (-7.12 pp). Subsequent GRPO yields small, statistically significant adjustments on 173 test utterances (1.69%), driving English script confusion (SCR_en) from 6.45% to 6.36% (p <= 0.001) and SPER to 28.70% (p = 0.022 clustered). Crucially, ablation reveals that in this SFT-converged encoder-decoder setting, an auxiliary script reward term yields performance identical to MER-only optimization because rollout candidate groups exhibit near-zero script variance on synthetic data. On monolingual benchmarks, SFT improves Nepali MER by 2.71 pp (stabilizing at 18.42% after GRPO, a 2.38 pp gain over baseline) while preserving English lexical WER (3.38% vs. 3.33%). Verified transcripts, timestamps, episode metadata, and reconstruction scripts are publicly released.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2610.30860 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-23T15:09:38.331Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).