Exploring the Effect of Segmentation and Vocabulary Size on Speech Tokenization for Speech Language Models
This work addresses the optimization of speech tokenization for speech language models, which is incremental as it fine-tunes existing methods for better efficiency and performance in spoken language understanding.
The paper investigates how segmentation width and vocabulary size affect speech tokenization for speech language models, finding that moderately coarse segmentation and larger cluster sizes improve performance, with the best model reducing training data by 50% and runtime by 70%.
The purpose of speech tokenization is to transform a speech signal into a sequence of discrete representations, serving as the foundation for speech language models (SLMs). While speech tokenization has many options, their effect on the performance of SLMs remains unclear. This paper investigates two key aspects of speech tokenization: the segmentation width and the cluster size of discrete units. First, we segment speech signals into fixed/variable widths and pooled representations. We then train K-means models in multiple cluster sizes. Through the evaluation on zero-shot spoken language understanding benchmarks, we find the positive effect of moderately coarse segmentation and bigger cluster size. Notably, among the best-performing models, the most efficient one achieves a 50% reduction in training data and a 70% decrease in training runtime. Our analysis highlights the importance of combining multiple tokens to enhance fine-grained spoken language understanding.