Overcoming Data Scarcity in Speaker Identification: Dataset Augmentation with Synthetic MFCCs via Character-level RNN

Mathematics

• Speaker Identification 100%
• Augmentation 57%
• Synthetic Data 38%
• Character 37%
• Speech 31%
• Speaker Recognition 16%
• Binary Classification 11%
• Data Augmentation 10%
• Recurrent Neural Networks 10%
• Large Data 10%
• Tuning 9%
• Classification Problems 8%
• Availability 8%
• Training 8%
• Neural Networks 7%
• Learning 7%
• Object 5%
• Metric 5%
• Class 4%
• Form 3%

Engineering & Materials Science

• Recurrent neural networks 16%
• Deep neural networks 15%
• Tuning 12%
• Availability 11%