ETRI-Knowledge Sharing Plaform

Self-Improving Isolated Word Recognition Compensation Method Artificial Intelligence System Acoustic Model Objective Function Parameter Reduction Barrier-Free Frame-Level Low-Resource Nonlinear Manifold Within-Class Covariance Normalization Universal Background Model Truncated Singular Value Decomposition Automatic Translation Distance Method Embedded Device Model Parameter Non-Native Speech Position Information Word Error Rate Similar Words Distance Algorithm Reduction Technique Speech Features Word-Level Vector Based Training Data Correct Position Channel Effect Factorization Approach Computing Technology Robust Speech Recognition Speech Recognition System System Performance Cosine Distance Channel Compensation Deep Neural Network(DNN) I-Vector Linear Discriminant Analysis(LDA) Matrix Factorization Manifold Structure Support Vectormachine(SVM) Phone Recognition Total Variability Word Similarity Non-Similar Automatic Speech Recognition(ASR) Best Performance Distance Measure Frequency Ratio Low-Rank Matrix Acoustic Features Euclidean Norm Target Matrix Confidence Measure Computational Complexity Cognitive Computing Decoding Method Language Barrier Manifold Regularized Speech Feature Vectors