Applying differentiable k-means to token-based ASR lets the tokenizer, the speech encoder, and the recognizer be optimized together, improving WER and making tokens more phoneme-like.
A lower value indicates the robustness of discrete tokens to non-linguistic information
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Differentiable K-means for Fully-optimized Discrete Token-based ASR
Applying differentiable k-means to token-based ASR lets the tokenizer, the speech encoder, and the recognizer be optimized together, improving WER and making tokens more phoneme-like.