{"slug": "dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning", "title": "Dissecting Sensitivity to Training Language in Self-Supervised Speech Learning Using Neural Audio Codec Tokens", "summary": "A systematic analysis of language sensitivity in codec-based self-supervised learning (SSL) models shows that downstream performance is insensitive to the neural audio codec (NAC) training language but strongly dependent on the SSL pre-training language, according to a new arXiv preprint (2607.26350v1). The findings suggest a single NAC can be reused across languages, while aligning the SSL pre-training language with the target language is crucial.", "body_md": "arXiv:2607.26350v1 Announce Type: cross\nAbstract: Neural audio codecs (NACs) have become popular for obtaining speech representations as discrete tokens. Beyond compression, discrete tokens can be used to train self-supervised learning (SSL) models. Such models, referred to as codec-based SSL models, reduce data storage and computational cost, enabling scalable SSL pre-training. However, their language sensitivity remains unclear. When the language changes, codec-based SSL models may require retraining, which undermines their efficiency. In this paper, we present a systematic analysis of language sensitivity by varying either the NAC training language or the SSL pre-training language while keeping the other fixed. Experimental results show that downstream performance is insensitive to the NAC training language but strongly dependent on the SSL pre-training language. These findings suggest that a single NAC can be reused across languages, while aligning the SSL pre-training language with the target language is crucial.", "url": "https://wpnews.pro/news/dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning", "canonical_source": "https://www.machinebrief.com/news/dissecting-sensitivity-to-training-language-in-self-supervis-nrc8", "published_at": "2026-07-30 04:00:00+00:00", "updated_at": "2026-07-30 08:01:55.286890+00:00", "lang": "en", "topics": ["machine-learning", "artificial-intelligence", "natural-language-processing"], "entities": ["arXiv"], "alternates": {"html": "https://wpnews.pro/news/dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning", "markdown": "https://wpnews.pro/news/dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning.md", "text": "https://wpnews.pro/news/dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning.txt", "jsonld": "https://wpnews.pro/news/dissecting-sensitivity-to-training-language-in-self-supervised-speech-learning.jsonld"}}