Robust Speaker Verification Using Deep Weight Space Ensemble

Keyphrases

• Robust Speaker Verification 100%
• Weighted Spaces 100%
• Text-dependent 75%
• Speaker Verification 75%
• Cross-channel 50%
• Model Fine-tuning 50%
• Domain Shift 50%
• Bilingual 25%
• Microphone 25%
• Error Surface 25%
• Low Error 25%
• Text-independent Speaker Verification 25%
• Multiple Models 25%
• Linear Path 25%
• Target Domain 25%
• Deep Neural Network 25%
• Bilingual Speakers 25%
• Telephone Speech 25%
• Speech Content 25%
• Regularization Effect 25%
• Challenging Problems 25%

Computer Science

• Speaker Verification 100%
• Deep Neural Network 40%
• Telephone 20%
• Regularization 20%

Physics

• Neural Network 100%