ETRI-Knowledge Sharing Plaform

Model Parameter Deep Neural Network(DNN) End Point Detection(EPD) Speech Signals Computational Complexity Word Error Rate Statistical Model Acoustic Features Model-Based Robust Speech Recognition Finite State Transducer Phone Recognition Automatic Speech Recognition(ASR) Point Of Interest Dnn-Based Acoustic Model Likelihood Ratio Test Non-Speech Performance Gain Inter-Frame Feature Vector Signal-To-Noise Speech Corpora Speech Corpus Two-Stage Taylor Series Speech Recognition Accuracy Front-End Convolution Neural Network(CNN) Recognition Performance CD-DNN-HMM Barrier-Free Synchronization Procedure Nonlinear Manifold Acoustic Source Behavior Strategy Conversational Speech Recognition Car Body Car Navigation System Dynamic Characteristics Detection Error Frame-Level Emotional Speech Filter Estimation Graph Regularized Human-Robot Interface Impact Noise Low-Resource Linear Structure Linear Separability Moving Vehicle Objective Function Parameter Reduction Performance Improvement Phone Error Rate Polynomial Expansion Reconstruction Method Raw Data Student Learning Speech Classification Wall Street Truncated Singular Value Decomposition Normalized Cross-Correlation Automatic Translation Cepstral Coefficients Cross-Correlation Function(CCF) Computationally Efficient Cost Function Decision Logic Decision Parameters Embedded Device Error Reduction Localization Method Large Vocabulary Level-Crossing Mechanical Sensing Mobile Environment Post-Processing Unlabeled Samples Wiener Filter Unified Framework Self-Training Speech Generation Spectrum Reconstruction Spectral Changes Policy Gradient Method Spatial Information Source Domain Estimation Method Accurate Estimation Background Noise Blind Source Separation(BSS) Context-Dependent Cascaded Decision Feedforward Neural Networks Handwritten Digit Recognition Korean Speech Noise Reduction(NR) Noise Suppression Pre-Trained Model Road Conditions Reduction Technique Unlabeled Data Real-Time Operation Room Acoustic Weight Matrix Training Method Semi-Supervised Training Scoring Model Hmm Based Speech Recognition Spectral Fluctuations Service Performance Speech Emotion Recognition Speech Quality Compensation Method Double-Delta Distributed Speech Recognition(DSR) Error Back-Propagation External Knowledge Hidden Markov Model High-Dimensional Features Input Signal Level Alignment Non-Stationary Noise Prior Knowledge Unsupervised Domain Adaptation Robot System Separation Device Spectral Characteristics Speech Recognition System Signal Distortion Signal Detection Target Location Factorization Approach Acoustic Model Training Artificial Intelligence System Conventional Methods Channel Selection Feature Extraction Method Graph-Based Approach Human Knowledge Localization Algorithm Learning Algorithms Mel-Frequency Cepstrum Coefficients(MFCC) Matrix Factorization Manifold Structure Processing Method Performance Degradation Viterbi Algorithm Target Domain Spontaneous Speech Speech Analysis Signal Detector Similarity Information Signal Noise Ratio(SNR) Time-Frequency Mask Synthesized Speech Nonlinear Activation Function Back Propagation Algorithm Car Interior Computational Efficiency Computational Cost Data-Driven Learning Engineering Approach Isolated Word Low-Rank Matrix L1 And L2-Norm Learning Methods Microphone Arrays Nonlinear Input Resource-Constrained Retail Market Sound Source Localization Statistical Hypothesis Testing Sub-Frame Pseudo Labels Zero-Crossing Voice Command Speech Separation Subjective Preference Nonlinear Function Automatic Speech Recognizer Dynamic Features Decision Knowledge Data Augmentation Digital Signal Processing Driving Conditions Euclidean Norm Knowledge-Based Local Linear Probabilistic Decision Parameter Optimization Posterior Probability Estimation Test Set Training Process Training Cycle Time-Domain Reinforcement Learning Algorithm Reduction Rate Spectrum Analyzer Speech Recordings Target Matrix Temporal Dynamics Computational Auditory Scene Analysis Confidence Measure Cross-Corpus Degraded Service Failure Rate Filter Bank Generation Process Input Features Language Barrier Manifold Regularized Noise Separation Noisy Speech Parameter Space Voice Recognition Vehicle Body Robust Recognition Speech Events 3-Level Self-Improving Time-Scale Modification(TSM)