Throughout this study, we used HuBERT [7] as an SSL model and extracted representations from the ninth layer

Tokenization Methods to be Explored As depicted in Figure 1, we first extract continuous speech representations of input speech using the SSL model

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

browse 1 citing papers

representative citing papers

cs.CL · 2025-05-23 · conditional · novelty 6.0

Using 80 ms speech segments and 16,384 sound tokens improves zero-shot spoken language understanding and cuts training cost by up to 70%.

Showing 1 of 1 citing paper.

Exploring the Effect of Segmentation and Vocabulary Size on Speech Tokenization for Speech Language Models cs.CL · 2025-05-23 · conditional · none · ref 2
Using 80 ms speech segments and 16,384 sound tokens improves zero-shot spoken language understanding and cuts training cost by up to 70%.