Quantization of SSL speech representations prioritizes segmental phonetics over suprasegmental lexical tone, reducing reliability for tone languages.
To use these continuous representations in downstream tasks, it is often necessary to discretise them into Discrete Speech Units (DSUs)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Lexical Tone is Hard to Quantize: Probing Discrete Speech Units in Mandarin and Yor\`ub\'a
Quantization of SSL speech representations prioritizes segmental phonetics over suprasegmental lexical tone, reducing reliability for tone languages.