ETRI-Knowledge Sharing Plaform

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