ETRI-Knowledge Sharing Plaform

Model Parameter Speech Signals Automatic Speech Recognition(ASR) Speech Recognition System Self-Improving Acoustic Features Artificial Intelligence System Data Augmentation Car Navigation System Wall Street Convolution Neural Network(CNN) Reduction Technique Korean Speech Feature Vector Phone Recognition Performance Degradation Point Of Interest Recognition Performance Training Data Text-Based End To End(E2E) Confidence Measure Computational Complexity CD-DNN-HMM Parameter Reduction Conversational Speech Recognition Phone Error Rate Assessment Method Student Learning Native Language Performance Improvement Conversion Model Error Reduction Word Error Rate Embedded Device Gaussian Mixture Model(GMM) Taylor Series Bidirectional Lstm Weight Matrix Semi-Supervised Training Speech Features Deep Learning(DL) Based System Compensation Method Linear Discriminant Analysis(LDA) Speech Database Large-Scale Recurrent Neural Network(RNN) Support Vectormachine(SVM) Target Domain I-Vector Mel-Frequency Cepstrum Coefficients(MFCC) Spontaneous Speech Learning Methods Training Process Performance Gain Reduction Rate Dnn-Based Acoustic Model Spectral Features Input Features Latent Variables Response Latency Augmentation Techniques Map-Based Raw Data Objective Function Paired Data Behavior Strategy GMM-HMM Within-Class Covariance Normalization Parameter Estimation Polynomial Expansion Universal Background Model Shared Task Frequency Filtering Healthy Controls Signal Characteristics Discriminative Information Nonlinear Manifold Sequential Data Single Parameter Linear Structure Maximum A Posteriori Spoken Language Understanding Second Language Barrier-Free Multiple Domains Filtering Method Viterbi Decoding Pattern Extraction Training Model High Accuracy Training Based Graph Regularized Data-Driven Approach Linear Separability Channel Normalization Context-Sensitive Low-Resource Speech-To-Text(STT) Priori Signal-To-Noise Ratio Truncated Singular Value Decomposition Smartphone-Based Moving Vehicle Noisy Environments Server-Client Speech Generation Source Domain Large Vocabulary Two-Stage Mental Illness User-Perceived Latency Cost Function Unlabeled Samples Band Pass Filter(BPF) Qualitative Characteristics Policy Gradient Method Image Recognition High Performance Decoding Algorithm Educational Use Automatic Translation Batch Processing Speech Recognition Accuracy Reduced Model Language Modeling Self-Training Front-End Linear Regression Neural Network Approach English Sentence Conversion Method Transformer-Based Learning Process Vector Based Long Sequence Scoring Model Fixed-Point Context-Dependent Pre-Trained Model English Proficiency Feed-Forward Deep Neural Network Real-World Applications Language Generation Isolated Word Recognition Embedding Model Spectral Domain Training Method Assessment System Data Sets Handwritten Digit Recognition Unlabeled Data Computer Assisted Language Learning Post-Processing Method Transfer Learning Classification Error Speech Corpus Speech Modeling Connectionist Temporal Classification(CTC) State-Of-The-Art Hmm Based Speech Recognition Word-Level Feedforward Neural Networks Major Depressive Disorder(MDD) Limited Resources Voice Activity Detector Low-Rank Matrix Approximation Generative Adversarial Network Level Alignment Factorization Approach Higher Performance Real-Time Intelligent Robot Diagonal Matrix Computing Technology Prior Knowledge External Knowledge Speech Translation Random Selection Optimal Weights Local Patches Acceptance Or Rejection Depressive Disorders Speech Quality Error Back-Propagation Inter-Frame Noise Components Speech Synthesis Channel Effect Korean Language Linguistic Content Sound Quality High-Dimensional Features Machine Learning Models Hidden Markov Model Human Knowledge Learning Algorithms N-Gram Lm Loss Function Output Snr Count-Based Learning Based Approach Output Prediction Spectral Peaks Feature Normalization Experimental Result Network-Based Linguistic Information Temporal Context Superior Performance Learning Network Single-Channel Speech Enhancement Late Fusion Graph-Based Approach Proficiency Evaluation Read Speech Proficiency Assessment Beam Search Acoustic Model Training Feature Extraction Method Latent Semantic Analysis(LSI) Frame Skipping Channel Compensation Learning System Total Variability Decision-Directed Approach Language Identification Real-World Data Manifold Structure Classification Accuracy Active Learning Method Speaker Conversion Signal Processing Auxiliary Networks Cosine Distance Acoustic Characteristics Matrix Factorization Speech Assessment Language Learners Signal Noise Ratio(SNR) Online Speech Recognition Low Latency Pseudo Labels Nonlinear Activation Function Back Propagation Algorithm Early Fusion Proposed Model Mental Health Treatment Phase Information Engineering Approach Art Performance Acoustic Properties Best Performance Ensemble Classifiers Quiz Game Integrated System Deep Convolutional Neural Networks Speaker Similarity Finite State Entropy Loss Retail Market L1 And L2-Norm Accuracy Improvement Nonlinear Input Dimension Reduction Self-Supervised Learning Latent Semantics Many-To-Many Style Conversion Data Sparseness Answer Extraction Piecewise Linear Nonlinear Function Research Topics Phase-Based Parameter Optimization Pronunciation Evaluation Window Size Magnitude And Phase Channel Noise Rule-Based Approach Target Matrix Speech Information Word Segmentation Euclidean Norm Reinforcement Learning Algorithm Chain Based Adaptive Modulation Training Cycle Test Set Local Linear Small Set Bidirectional Long Short-Term Memory Machine Learning Technique(MLT) Automatic Speech Recognizer Posterior Probability Estimation Normalization Method System Accuracy Driving Conditions Time-Domain Minimum Mean Square Error(MMSE) Peaks And Valleys Stable Model Case Studies Frequency Cepstral Coefficients Log Data Non-Native Speakers Feature Construction Filter Bank Language Barrier Noisy Speech Computation Method Question Answering Modulation Frequency Speech Feature Extraction Cnn Model Quality Enhancement Frame Delay Sparsity Constraint Generation Process Multi-Stage Manifold Regularized Linear Interpolation Automatic Diagnosis Second-Order Local Minima Classification Performance Rank Selection Differential Diagnosis Automatic Identification Text Data Cognitive Computing Depression Detection Priori Snr Efficient Computation Cross Entropy Parameter Space