Unsupervised Word Segmentation Using Temporal Gradient Pseudo-Labels
This work addresses the problem of segmenting speech into words without supervision, which is incremental as it builds on existing deep self-supervised features.
The paper tackles unsupervised word segmentation in audio by proposing a pseudo-labeling strategy based on temporal gradient magnitude of embeddings, achieving significant performance improvements over previous methods on two datasets.
Unsupervised word segmentation in audio utterances is challenging as, in speech, there is typically no gap between words. In a preliminary experiment, we show that recent deep self-supervised features are very effective for word segmentation but require supervision for training the classification head. To extend their effectiveness to unsupervised word segmentation, we propose a pseudo-labeling strategy. Our approach relies on the observation that the temporal gradient magnitude of the embeddings (i.e. the distance between the embeddings of subsequent frames) is typically minimal far from the boundaries and higher nearer the boundaries. We use a thresholding function on the temporal gradient magnitude to define a psuedo-label for wordness. We train a linear classifier, mapping the embedding of a single frame to the pseudo-label. Finally, we use the classifier score to predict whether a frame is a word or a boundary. In an empirical investigation, our method, despite its simplicity and fast run time, is shown to significantly outperform all previous methods on two datasets.