2020/07/01 by Bowen Shi, Shane Settle, Shi, Bowen +3
Computer Science · Engineering · Mathematics · #Algorithm #Artificial intelligence #Audio and Speech Processing (eess.AS) #Computation and Language (cs.CL) #Computer science #Decoding methods #FOS: Computer and information sciences #FOS: Electrical engineering #Feature (linguistics) #Feature vector #Hidden Markov model #Linguistics #Machine learning #Margin (machine learning) #Mathematics #Music and Audio Processing #Natural Language Processing Techniques #Natural language processing #Pattern recognition (psychology) #Sound (cs.SD) #Speech Recognition and Synthesis #Speech recognition #Viterbi algorithm #Viterbi decoder #Vocabulary #Word (group theory) #Word error rate #Word recognition #cs.CL #cs.SD #eess.AS #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2007.00183
SLT 2021
openalex publication_date 2020/07/01 · arxiv created 2020/11/24 · arxiv updated 2020/11/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06
Segmental models are sequence prediction models in which scores of hypotheses are based on entire variable-length segments of frames. We consider segmental models for whole-word ("acoustic-to-word") speech recognition, with the feature vectors defined using vector embeddings of segments. Such models are computationally challenging as the number of paths is proportional to the vocabulary size, which can be orders of magnitude larger than when using subword units like phones. We describe an efficient approach for end-to-end whole-word segmental models, with forward-backward and Viterbi decoding performed on a GPU and a simple segment scoring function that reduces space complexity. In addition, we investigate the use of pre-training via jointly trained acoustic word embeddings (AWEs) and acoustically grounded word embeddings (AGWEs) of written word labels. We find that word error rate can be reduced by a large margin by pre-training the acoustic segment representation with AWEs, and additional (smaller) gains can be obtained by pre-training the word prediction layer with AGWEs. Our final models improve over prior A2W models.