ETRI-Knowledge Sharing Plaform

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