ETRI-Knowledge Sharing Plaform

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