CONTRASTIVE SPEAKER EMBEDDING WITH SEQUENTIAL DISENTANGLEMENT

Keyphrases

• Disentanglement 100%
• Speaker Embedding 100%
• Speaker Identity 66%
• Contrastive Learning 66%
• Linguistic Content 66%
• Content Factors 66%
• Speaker Factor 66%
• Sequential Variational Autoencoder 66%
• SimCLR 66%
• Speech Signal 33%
• Learning Framework 33%
• Speech Segments 33%
• Embedding Space 33%
• Contrastive Loss 33%
• Discriminative Embedding 33%
• VoxCeleb 33%

Computer Science

• Contrastive Learning 100%
• Variational Autoencoder 100%
• Experimental Result 50%
• Learning Framework 50%
• Contrastive Loss 50%
• Embedding Space 50%
• Invariant 50%