Large vocabulary continuous speech recognition using WFST-based linear classifier for structured data

Arts & Humanities

• Acoustics 11%
• Classifier 72%
• Continuous Speech 95%
• Conventional 9%
• Decoding 56%
• Graph 12%
• Hidden Markov Model (HMM) 16%
• N-gram 17%
• Performance 4%
• Phoneme 13%
• Speech Recognition 100%
• Telephone 12%
• Vocabulary 74%

Medicine & Life Sciences

• Acoustics 15%
• Computer Neural Networks 20%
• Linear Models 24%
• Psychology Recognition 54%
• Transducers 62%
• Vocabulary 76%

Social Sciences

• acoustics 16%
• linear model 34%
• training method 18%
• vocabulary 49%