Maximal figure-of-merit framework to detect multi-label phonetic features for spoken language recognition

Keyphrases

• Phonetic Features 100%
• Bottleneck Features 100%
• Spoken Language Recognition 100%
• Maximal Figure-of-merit 100%
• Language Recognition 75%
• Spectral Features 50%
• Deep Neural Network 50%
• Speech Attributes 50%
• Speech-language 25%
• Recognition System 25%
• Architecture-centric 25%
• Output Layer 25%
• Recognition Accuracy 25%
• Spoken Language 25%
• Backpropagation 25%
• High Dimension 25%
• Learning Objectives 25%
• Intrinsic Relationship 25%
• Articulatory Features 25%
• Place of Articulation 25%
• Manner of Articulation 25%
• Speech Segments 25%
• Evaluation Challenges 25%
• Phonetics 25%
• Deep Learning Architectures 25%
• Deep Learning Framework 25%
• Feature Attribute 25%
• Speech Features 25%
• Systems-based 25%
• Training Scheme 25%
• Differentiable Measure 25%

Computer Science

• Spoken Language 100%
• Spectral Feature 66%
• Deep Neural Network 66%
• High Dimensionality 33%
• Deep Architecture 33%
• Deep Learning Method 33%
• Learning Framework 33%
• Postprocessing 33%
• Experimental Evidence 33%
• Recognition System 33%
• Learning Target 33%
• Recognition Accuracy 33%