5 papers
Alternating Weak Triphone/BPE Alignment Supervision from Hybrid Model Improves End-to-End ASR
Jintao Jiang, Yingbo Gao, Mohammad Zeineldeen +1
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 usi…
Take the Hint: Improving Arabic Diacritization with Partially-Diacritized Text
Parnia Bahar, Mattia Di Gangi, Nick Rossenbach +1
Automatic Arabic diacritization is useful in many applications, ranging from reading support for language learners to accurate pronunciation predictor for downstream tasks like spe…
Improving Language Model Integration for Neural Machine Translation
Christian Herold, Yingbo Gao, Mohammad Zeineldeen +1
The integration of language models for neural machine translation has been extensively studied in the past. It has been shown that an external language model, trained on additional…
Robust Knowledge Distillation from RNN-T Models With Noisy Training Labels Using Full-Sum Loss
Mohammad Zeineldeen, Kartik Audhkhasi, Murali Karthick Baskar +1
This work studies knowledge distillation (KD) and addresses its constraints for recurrent neural network transducer (RNN-T) models. In hard distillation, a teacher model transcribe…
Improving the Training Recipe for a Robust Conformer-based Hybrid Model
Mohammad Zeineldeen, Jingjing Xu, Christoph Lüscher +2
Speaker adaptation is important to build robust automatic speech recognition (ASR) systems. In this work, we investigate various methods for speaker adaptive training (SAT) based o…