Pith. sign in

REVIEW 3 cited by

Simple and Effective Zero-shot Cross-lingual Phoneme Recognition

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2109.11680 v1 pith:6ZAS3DOG submitted 2021-09-23 cs.CL cs.LGcs.SD

Simple and Effective Zero-shot Cross-lingual Phoneme Recognition

classification cs.CL cs.LGcs.SD
keywords languagescross-lingualdatalabeledlearningmodelpretrainedrecognition
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Recent progress in self-training, self-supervised pretraining and unsupervised learning enabled well performing speech recognition systems without any labeled data. However, in many cases there is labeled data available for related languages which is not utilized by these methods. This paper extends previous work on zero-shot cross-lingual transfer learning by fine-tuning a multilingually pretrained wav2vec 2.0 model to transcribe unseen languages. This is done by mapping phonemes of the training languages to the target language using articulatory features. Experiments show that this simple method significantly outperforms prior work which introduced task-specific architectures and used only part of a monolingually pretrained model.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Multilingual Phonological Feature Recognition with Self-Supervised Speech Models

    cs.CL 2026-05 unverdicted novelty 7.0

    PhonoQ-2.0 directly predicts structured phonological features from self-supervised models with a gating mechanism, outperforming phoneme baselines by 8+ macro-F1 points on average across in-domain, out-of-domain, and ...

  2. Harf-Speech: A Clinically Aligned Framework for Arabic Phoneme-Level Speech Assessment

    eess.AS 2026-03 unverdicted novelty 6.0

    Harf-Speech delivers phoneme-level Arabic pronunciation scores that correlate 0.79 with certified speech-language pathologists on 40 utterances.

  3. Harf-Speech: A Clinically Aligned Framework for Arabic Phoneme-Level Speech Assessment

    eess.AS 2026-03 conditional novelty 5.0

    A modular Arabic phoneme-level pronunciation scorer reaches Pearson 0.791 and ICC 0.659 with three SLPs on 40 utterances, with best speech-to-phoneme PER of 8.92%.