method · first seen 6 Nov 2026
acoustic word embeddings
Also: AWE
Fixed-dimensional vector representations of spoken word segments capturing lexical content from audio.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Asymmetric Correspondence Learning for Lexical and Style Embeddings: Interpreting Latent Space via Acoustic FeaturesPapers · 6 Nov 2026 · Michael Whealing, Robert Gaizauskas, Thomas Hain
Latest sources
- Asymmetric Correspondence Learning for Lexical and Style Embeddings: Interpreting Latent Space via Acoustic FeaturesPapers · 6 Nov 2026 · Michael Whealing, Robert Gaizauskas, Thomas Hain