DiscoPhon: Benchmarking the Unsupervised Discovery of Phoneme Inventories With Discrete Speech Units
This provides a standardized benchmark for researchers in speech processing to assess unsupervised phoneme discovery methods, but it is incremental as it builds on existing models and focuses on evaluation rather than new techniques.
The authors tackled the problem of evaluating unsupervised phoneme discovery from discrete speech units by introducing DiscoPhon, a multilingual benchmark covering 12 languages, and found that current models like HuBERT and SpidR produce units that correlate well with phonemes, though performance varies across languages.
We introduce DiscoPhon, a multilingual benchmark for evaluating unsupervised phoneme discovery from discrete speech units. DiscoPhon covers 6 dev and 6 test languages, chosen to span a wide range of phonemic contrasts. Given only 10 hours of speech in a previously unseen language, systems must produce discrete units that are mapped to a predefined phoneme inventory, through either a many-to-one or a one-to-one assignment. The resulting sequences are evaluated for unit quality, recognition and segmentation. We provide four pretrained multilingual HuBERT and SpidR baselines, and show that phonemic information is available enough in current models for derived units to correlate well with phonemes, though with variations across languages.