Alternating Weak Triphone/BPE Alignment Supervision from Hybrid Model Improves End-to-End ASR

arXiv:2402.15594v1 Announce Type: new
Abstract: In this paper, alternating weak triphone/BPE alignment supervision is proposed to improve end-to-end model training. Towards this end, triphone and BPE alignments are extracted using a pre-existing hybrid ASR system. Then, regularization effect is obtained by cross-entropy based intermediate auxiliary losses computed on such alignments at a mid-layer representation of the encoder for triphone alignments and at the encoder for BPE alignments. Weak supervision is achieved through strong label smoothing with parameter of 0.5. Experimental …

abstract alignment arxiv asr cross-entropy cs.cl cs.sd eess.as entropy hybrid intermediate losses paper regularization supervision training type