Search papers, labs, and topics across Lattice.
6
0
5
2
A DTW-based framework using self-supervised representations outperforms human raters in assessing L2 phonetic accuracy and rhythm, challenging traditional assessment methods.
Manipulating specific dimensions in speech feature subspaces can enable precise control over characteristics like pitch and intensity, revolutionizing speech synthesis techniques.
Achieving spoken word segmentation without any text supervision, this method outperforms traditional neural approaches while enhancing interpretability.
Graph-based clustering can recover Zipfian distributions in unsupervised term discovery, outperforming traditional K-means methods.
Current lexicon evaluation methods are biased towards large clusters, but new metrics reveal a more accurate picture of lexicon quality.
Unlocking control over speaker characteristics in speech synthesis is as simple as tweaking individual dimensions of self-supervised speech features.
