ETRI-Knowledge Sharing Plaform

Model Parameter Semi-Supervised Training Acoustic Features Phone Recognition Frame-Level Wall Street Self-Improving Reduction Technique Feature Vector Robust Speech Recognition Speech Recognition System Artificial Intelligence System Finite State Transducer Automatic Speech Recognition(ASR) Speech Signals Statistical Hypothesis Testing Decision Knowledge Dnn-Based Acoustic Model Training Process Knowledge-Based Non-Speech Confidence Measure Failure Rate Input Features Compensation Method Speech Corpora End Point Detection(EPD) Weight Matrix Cascaded Decision Self-Training Word Error Rate Unified Framework Taylor Series Decision Parameters Decision Logic Convolution Neural Network(CNN) Phone Error Rate Parameter Reduction CD-DNN-HMM Data-Driven Car Navigation System Dynamic Characteristics Discriminative Information Behavior Strategy Truncated Singular Value Decomposition Universal Background Model Speech-To-Text(STT) Linear Separability Within-Class Covariance Normalization Embedded Devices Graph Regularized Low-Resource Linear Structure Nonlinear Manifold Objective Function Polynomial Expansion Paired Data Speech Classification Raw Data Single Parameter Student Learning Cost Function Unlabeled Samples Error Reduction Non-Native Speech Post-Processing Policy Gradient Method Speech Recognition Accuracy Spectral Changes Reduced Model Speech Generation Training Data Classification Error Training Method Transfer Learning Vector Based Unlabeled Data Word-Level Feedforward Neural Networks Feed-Forward Deep Neural Network Handwritten Digit Recognition Hmm Based Speech Recognition Isolated Word Recognition Pre-Trained Model Spectral Fluctuations Speech Corpus Scoring Model Channel Effect Computing Technology Double-Delta Diagonal Matrix External Knowledge Error Back-Propagation Factorization Approach Hidden Markov Model High-Dimensional Features Inter-Frame Low-Rank Matrix Approximation Model-Based Non-Stationary Prior Knowledge Total Variability Auxiliary Networks Acoustic Model Training Cosine Distance Channel Compensation Cross-Entropy Classification Accuracy Temporal Context Feature Extraction Method Graph-Based Approach Human Knowledge I-Vector Linear Discriminant Analysis(LDA) Loss Function Learning Based Approach Learning Algorithms Matrix Factorization Manifold Structure Network-Based Processing Method Performance Degradation Speech Analysis Support Vectormachine(SVM) Spontaneous Speech Signal Processing Accuracy Improvement Best Performance Backpropagation Algorithm Sub-Frame Entropy Loss Engineering Approach L1 And L2-Norm Nonlinear Activation Function Nonlinear Input Pseudo Labels Temporal Dynamics Dynamic Features Data Augmentation Training Cycle Time-Domain Target Matrix Test Set Window Size End-To-End Asr Euclidean Norm Local Linear Nonlinear Function Probabilistic Decision Posterior Probability Estimation Performance Gain Parameter Optimization Spectrum Analyzer Reduction Rate Reinforcement Learning Algorithm Speech Information Computational Complexity Cognitive Computing Empirical Study Generation Process Feature Construction Classification Performance Manifold Regularized Parameter Space Recognition Performance Rank Selection Sparsity Constraint