Acoustically Grounded Word Embeddings for Improved Acoustics-to-word Speech Recognition.

Shane Settle, Kartik Audhkhasi, Karen Livescu, Michael Picheny

Browse the full ICASSP paper archive.