Self-Supervised Audio-Visual Speaker Representation with Co-Meta Learning

Keyphrases

• Audio 100%
• Meta-learning 100%
• Speaker Representation 100%
• Pseudo Label 66%
• Learning Framework 33%
• Relative Improvement 33%
• Network Parameters 33%
• Co-teaching 33%
• Visual Learning 33%
• VoxCeleb 33%
• Evaluation Dataset 33%
• Supervisory Signal 33%
• Self-supervised Learning 33%
• Hard Samples 33%
• Self-supervised Speaker Verification 33%
• Gradient Regularization 33%
• Audio Representation Learning 33%

Computer Science

• Meta-Learning 100%
• Massive Amount 33%
• Learning Framework 33%
• Speaker Verification 33%
• Visual Representation 33%
• Representation Learning 33%
• Self-Supervised Learning 33%
• Regularization 33%
• Network Parameter 33%