Attention-based multi-task learning for speech-enhancement and speaker-identification in multi-speaker dialogue scenario

Keyphrases

• Speech Enhancement 100%
• Speaker Identification 100%
• Attention-based 100%
• Multi-speaker 100%
• Multi-task Learning 100%
• Deep Neural Network 42%
• Attention Net 42%
• Memory Models 14%
• Learning Mechanisms 14%
• Noisy Environment 14%
• Evaluation Results 14%
• Learning Approaches 14%
• Attention Mechanism 14%
• Speech Signal 14%
• Speech Quality 14%
• Noisy Input 14%
• Speech Intelligibility 14%
• Taiwan Mandarin 14%
• Weight-based 14%
• Learning Structure 14%
• Speaker Identity 14%
• Attention Weight 14%
• Weighting Mechanism 14%
• Test Sentence 14%
• Hearing in Noise Test 14%
• Robust Acoustic Features 14%
• Multi-model Learning 14%
• Task Attention 14%

Computer Science

• Speaker Identification 100%
• Speech Enhancement 100%
• Multitask Learning 100%
• Deep Neural Network 42%
• Long Short-Term Memory Network 42%
• Noisy Environment 14%
• Neural Network Model 14%
• Evaluation Result 14%
• Learning Approach 14%
• Attention (Machine Learning) 14%
• Acoustic Feature 14%
• structure learning 14%
• Memory Model 14%

Engineering

• Speech Enhancement 100%
• Speaker Identification 100%
• Learning Task 100%
• Long Short-Term Memory 42%
• Deep Neural Network 42%
• Specifies 14%
• Network Model 14%
• Acoustic Feature 14%
• Speech Signal 14%
• Noisy Environment 14%
• Learning Approach 14%
• Related Task 14%
• Noise Test 14%