Exploring hybrid ctc/attention end-to-end speech recognition with gaussian processes

Keyphrases

• Attention Mechanism 66%
• Attention Model 16%
• Attention Network 16%
• Beam Search 33%
• Character Error Rate 66%
• Connectionist Temporal Classification 100%
• Degrees of Freedom 16%
• End-to-end Speech Recognition 100%
• Evidence-based 16%
• Feature Sequence 16%
• Gaussian Optimization 16%
• Gaussian Process 100%
• Hybrid Architecture 16%
• Language Model 50%
• Model Feedback 16%
• Model Weight 16%
• Network Language 16%
• Network Parameters 16%
• Parameter Configuration 16%
• RNNLM 66%
• Sequential Alignment 16%
• Speech Features 16%
• Word Error Rate 50%

Computer Science

• Attention (Machine Learning) 100%
• Configuration Parameter 20%
• Hybrid Architecture 20%
• Language Modeling 60%
• Network Parameter 20%
• Process Optimization 20%
• Speech Recognition 100%
• Word Error Rate 60%