3 papers
cs.LG2021
EDDA: Explanation-driven Data Augmentation to Improve Explanation Faithfulness
Ruiwen Li, Zhibo Zhang, Jiani Li +5
Recent years have seen the introduction of a range of methods for post-hoc explainability of image classifier predictions. However, these post-hoc explanations may not always be fa…
cs.SD2018
Quaternion Convolutional Neural Networks for End-to-End Automatic Speech Recognition
Titouan Parcollet, Ying Zhang, Mohamed Morchid +4
Recently, the connectionist temporal classification (CTC) model coupled with recurrent (RNN) or convolutional neural networks (CNN), made it easier to train speech recognition syst…
stat.ML2018
Quaternion Recurrent Neural Networks
Titouan Parcollet, Mirco Ravanelli, Mohamed Morchid +4
Recurrent neural networks (RNNs) are powerful architectures to model sequential data, due to their capability to learn short and long-term dependencies between the basic elements o…