PulseAugur
EN
LIVE 04:47:09

Cross-lingual transfer learning shows mixed results for low-resource ASR

By PulseAugur Editorial · [4 sources] ·

Researchers have explored cross-lingual transfer learning to improve automatic speech recognition (ASR) for low-resource languages. One study successfully used Sinhala to enhance Dhivehi ASR, achieving a 12.89% Word Error Rate (WER) through continual pre-training and fine-tuning. In contrast, another study found that for large multilingual ASR models, pre-adaptation on related auxiliary languages did not yield significant improvements for low-resource African languages, suggesting linguistic relatedness alone may not be sufficient in such contexts. AI

IMPACT Investigates methods to improve speech recognition for under-resourced languages, potentially broadening AI accessibility.

RANK_REASON The cluster contains two research papers on arXiv discussing cross-lingual transfer learning for automatic speech recognition.

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 4 sources. How we write summaries →

Cross-lingual transfer learning shows mixed results for low-resource ASR

COVERAGE [4]

  1. arXiv cs.CL TIER_1 English(EN) · Lukmal Ilyas, Nevidu Jayatilleke ·

    From Sinhala to Dhivehi: Cross-Lingual Transfer Learning for Low-Resource Speech Recognition

    arXiv:2607.06289v1 Announce Type: new Abstract: Dhivehi, the national language of the Maldives, is currently under-resourced for automatic speech recognition (ASR) and other NLP tasks. This study investigates whether cross-lingual transfer learning from Sinhala, a linguistically …

  2. arXiv cs.CL TIER_1 English(EN) · Nevidu Jayatilleke ·

    From Sinhala to Dhivehi: Cross-Lingual Transfer Learning for Low-Resource Speech Recognition

    Dhivehi, the national language of the Maldives, is currently under-resourced for automatic speech recognition (ASR) and other NLP tasks. This study investigates whether cross-lingual transfer learning from Sinhala, a linguistically related, relatively well-resourced Insular Indo-…

  3. arXiv cs.AI TIER_1 English(EN) · Andrei Florian, Cynthia Jayne Amol, Hope Kerubo Ombaba, Xiaoyu Cui, Boniface Mwau, Biatus Maina Kamau, Lilian Diana Awuor Wanzare, Christiane Fellbaum, Happy Buzaaba ·

    Evaluating the Effect of Linguistic Relatedness on Cross-Lingual Transfer in Large Multilingual Automatic Speech Recognition

    arXiv:2607.04814v1 Announce Type: cross Abstract: Extending automatic speech recognition (ASR) to low-resource African languages is constrained by the prohibitive demands of data collection at scale. A promising direction is to leverage linguistic relatedness to enhance cross-lin…

  4. arXiv cs.AI TIER_1 English(EN) · Happy Buzaaba ·

    Evaluating the Effect of Linguistic Relatedness on Cross-Lingual Transfer in Large Multilingual Automatic Speech Recognition

    Extending automatic speech recognition (ASR) to low-resource African languages is constrained by the prohibitive demands of data collection at scale. A promising direction is to leverage linguistic relatedness to enhance cross-lingual transfer from a related auxiliary language to…

RELATED ENTITIES

RELATED TOPICS