RNN-T For Latency Controlled ASR With Improved Beam Search
Neural transducer-based systems such as RNN Transducers (RNN-T) for automatic speech recognition (ASR) blend the individual components of a traditional hybrid ASR systems (acoustic model, language model, punctuation model, inverse text normalization) into one single model. This greatly simplifies training and inference and hence makes RNN-T a desirable choice for ASR systems. In this work, we investigate use of RNN-T in applications that require a tune-able latency budget during inference time. We also improved the decoding speed of the originally proposed RNN-T beam search algorithm. We evaluated our proposed system on English videos ASR dataset and show that neural RNN-T models can achieve comparable WER and better computational efficiency compared to a well tuned hybrid ASR baseline.
Code (0)
등록된 구현이 없습니다.
Tasks
Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Computational EfficiencyLanguage ModelingLanguage Modellingspeech-recognitionSpeech RecognitionText NormalizationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Radar Enhanced Multi-Armed Bandit for Rapid Beam Selection in Millimeter Wave Communications
Multi-arm bandit (MAB) algorithms have been used to learn optimal beams for millimeter wave communication systems. Here, the complexity of learning the optimal beam linearly scales with the number of beams, leading to hi…
Phased Array With Improved Beamforming Capability via Use of Double Phase Shifters
The passive electronically scanned array (PESA) is widely used due to its simple structure and low cost. {Its antenna weights have unit modulus and thus, only the weights phases can be controlled. PESA has limited degree…
Enhancing User Throughput in Multi-panel mmWave Radio Access Networks for Beam-based MU-MIMO Using a DRL Method
Millimeter-wave (mmWave) communication systems, particularly those leveraging multi-user multiple-input and multiple-output (MU-MIMO) with hybrid beamforming, face challenges in optimizing user throughput and minimizing …
Reinforcement LearningFast mmwave Beam Alignment via Correlated Bandit Learning
Beam alignment (BA) is to ensure the transmitter and receiver beams are accurately aligned to establish a reliable communication link in millimeter-wave (mmwave) systems. Existing BA methods search the entire beam space …
Incremental Blockwise Beam Search for Simultaneous Speech Translation with Controllable Quality-Latency Tradeoff
Blockwise self-attentional encoder models have recently emerged as one promising end-to-end approach to simultaneous speech translation. These models employ a blockwise beam search with hypothesis reliability scoring to …
Translation