Publication | Closed Access

Modeling Time-Frequency Patterns with LSTM vs. Convolutional Architectures for LVCSR Tasks

Sequence ModellingEngineeringMachine LearningPattern RecognitionTime-frequency PatternsComputer EngineeringLvcsr TasksTemporal Pattern RecognitionSpeech ProcessingComputer ScienceTimefrequency AnalysisDeep LearningReal-time LanguageRecurrent Neural NetworkConvolutional ArchitecturesSpeech Recognition