H Kheddar, M Hemis, Y Himeur - Information Fusion, 2024 - Elsevier
Recent advancements in deep learning (DL) have posed a significant challenge for automatic speech recognition (ASR). ASR relies on extensive training datasets, including …
Y Xiao, L Wu, J Guo, J Li, M Zhang… - IEEE Transactions on …, 2023 - ieeexplore.ieee.org
Non-autoregressive (NAR) generation, which is first proposed in neural machine translation (NMT) to speed up inference, has attracted much attention in both machine learning and …
Transformers have recently dominated the ASR field. Although able to yield good performance, they involve an autoregressive (AR) decoder to generate tokens one by one …
This paper proposes a method to relax the conditional independence assumption of connectionist temporal classification (CTC)-based automatic speech recognition (ASR) …
Fifth-generation (5G) wireless networks will likely offer high data rates, increased reliability, and low delay for mobile, personal, and local area networks. Along with the rapid growth of …
Abstract Fifth-generation (5G) wireless networks are likely to offer high data rates, increased reliability, and low delay for mobile, personal, and local area networks. Along with the rapid …
While Transformers have achieved promising results in end-to-end (E2E) automatic speech recognition (ASR), their autoregressive (AR) structure becomes a bottleneck for speeding up …
Recently, end-to-end models have been widely used in automatic speech recognition (ASR) systems. Two of the most representative approaches are connectionist temporal …