Multi-task learning using mismatched transcription for under-resourced speech recognition

Keyphrases

• Multi-task Learning 100%
• Mismatched Transcription 100%
• Under-resourced Speech Recognition 100%
• Multi-task Learning Framework 40%
• Nonsense 20%
• Audio 20%
• Labeled Data 20%
• Native Speaker 20%
• Acoustic Modeling 20%
• Mandarin 20%
• Large Set 20%
• One Solution 20%
• Adaptation Techniques 20%
• Modeling Performance 20%
• Transcriber 20%
• Under-resourced Languages 20%
• Georgian 20%
• IARPA 20%
• Babel 20%
• DNN Model 20%

Computer Science

• Speech Recognition 100%
• Multitask Learning 100%
• Learning Framework 66%
• Performance Model 33%
• Deep Neural Network 33%