Feature-space structural MAPLR with regression tree-based multiple transformation matrices for DNN

Mathematics

• Regression Tree 100%
• Feature Space 93%
• Transformation Matrix 87%
• Neural Networks 64%
• Linear regression 47%
• Decoding 32%
• Gaussian Mixture Model 28%
• Maximum Likelihood 25%
• Maximum a Posteriori 25%
• MAP Estimation 16%
• Generative Models 13%
• Alignment 11%
• Framework 11%
• Acoustics 10%
• Weighted Sums 10%
• Prior distribution 9%
• Experimental Results 9%
• Multiplication 7%
• Tend 7%
• Transform 7%

Engineering & Materials Science

• Linear regression 68%
• Deep neural networks 66%
• Maximum likelihood 38%
• Decoding 34%
• Acoustics 11%