Entity · paper

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

paperactiveprovisionalfrom-sinhala-to-dhivehi-cross-lingual-transfer-learning-for-low-resource-speech-recognition-5294ed53·1 events·first seen 5d ago

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

Co-occurring entities

KenLM

More like this (12)

IWSLT 2026 Cross-Lingual Voice Cloning Leveraging Audio-LLMs to Filter Speech-to-Speech Training Data Same Lesson, Different Story: Cross-Lingual Reconstruction of Cultural Narratives in Large Language Models Interleaved Speech Language Models Latently Work In Text Learning to Hear Hesitation: Continual Learning for Disfluency-Aware ASR When English Isn't the Best Teacher: Source Language Effects in Cross-Lingual In-Context Learning Echoes Across Vietnam's Highlands, Delta, and Coast: A Multilingual Corpus for Cham, Khmer, and Tay-Nung Speaker Group Encoding in Self-supervised Speech Recognition Models CapSpeech-TTS Data Synthesis and Parameter-Efficient Fine-Tuning for Low-Resource NMT: A Case Study on Q'eqchi' Mayan simultaneous speech-to-text translation Reasoning over Grammar: Can Synthetic Linguistic Reasoning Traces Enhance Low-Resource Machine Translation?

Recent events (1)

3arXiv · cs.CL·5d ago·source ↗

Cross-lingual transfer from Sinhala improves low-resource Dhivehi ASR

A new arXiv paper investigates cross-lingual transfer learning from Sinhala to improve automatic speech recognition for Dhivehi, a severely low-resource language of the Maldives. Across 17 experiments spanning five transfer paradigms, the best system—continual pre-training on Sinhala followed by Dhivehi fine-tuning with KenLM—achieves 12.89% WER, outperforming the Dhivehi-only baseline by 13.50% WER. A Turkish control experiment confirms that linguistic relatedness, not just additional data, drives the improvement.

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