Multi-microphone speech recognition integrating beamforming, robust feature extraction, and advanced DNN/RNN backend

Mathematics

• Beamforming 100%
• Speech Recognition 95%
• Feature Extraction 93%
• Recurrent Neural Networks 80%
• Neural Networks 58%
• Training 39%
• Speech Enhancement 23%
• Acoustic Model 20%
• Language Model 19%
• Speech 16%
• Memory Term 16%
• Enhancement 14%
• Network Model 14%
• Decoding 13%
• Baseline 13%
• Benchmark 12%
• Presentation 10%
• Methodology 9%
• Model 8%
• Generalization 6%
• Coefficient 6%

Engineering & Materials Science

• Microphones 65%
• Recurrent neural networks 65%
• Beamforming 61%
• Deep neural networks 60%
• Speech recognition 58%
• Feature extraction 48%
• Speech enhancement 15%
• Long short-term memory 14%
• Decoding 10%
• Acoustics 10%
• Computer simulation 6%