ICASSP 2022

Alignment-Learning based single-step decoding for accurate and fast non-autoregressive speech recognition

HYBRID RNN-T/ATTENTION-BASED STREAMING ASR WITH TRIGGERED CHUNKWISE ATTENTION AND DUAL INTERNAL LANGUAGE MODEL INTEGRATION

On Language Model Integration for RNN Transducer based Speech Recognition

IMPROVING PSEUDO-LABEL TRAINING FOR END-TO-END SPEECH RECOGNITION USING GRADIENT MASK

IMPROVING CTC-BASED SPEECH RECOGNITION VIA KNOWLEDGE TRANSFERRING FROM PRE-TRAINED LANGUAGE MODELS

PHONE-INFORMED REFINEMENT OF SYNTHESIZED MEL SPECTROGRAM FOR DATA AUGMENTATION IN SPEECH RECOGNITION

UNSUPERVISED DATA SELECTION FOR SPEECH RECOGNITION WITH CONTRASTIVE LOSS RATIOS

IMPORTANTAUG: A DATA AUGMENTATION AGENT FOR SPEECH

AISHELL-NER: NAMED ENTITY RECOGNITION FROM CHINESE SPEECH

JOINT SPEECH RECOGNITION AND AUDIO CAPTIONING

SRU++: PIONEERING FAST RECURRENCE WITH ATTENTION FOR SPEECH RECOGNITION

LATTENTION: LATTICE-ATTENTION IN ASR RESCORING