Interpreting and Explaining Deep Neural Networks for Classification of Audio Signals

By Sören Becker, Marcel Ackermann, Sebastian Lapuschkin, Klaus‐Robert Müller , Wojciech Samek. 2018

152 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Survey on Explainable Artificial Intelligence (XAI): Toward Medical XAI (2019, background): “Note that, so far, there is neither explanation nor interpretation of the inner working of an ML algorithm.”
Unmasking Clever Hans predictors and assessing what machines really learn (2019, methodology): “Explanation methods were successfully applied to a wide set of complex real-world problems such as the analysis of faces [81, 123, 124], EEG and fMRI data [23, 50], human gait [48], videos [125], speech [126], reinforcement learning [25], biological data [127], text [26,128,129], or comparing human ”
A Wholistic View of Continual Learning with Deep Neural Networks: Forgotten Lessons and the Bridge to Active and Open World Learning (2020, methodology): “We conduct experiments on two datasets: the CIFAR100 and the AudioMNIST (Becker et al., 2018) dataset.”
Explaining and Interpreting LSTMs (2019, methodology): “Further, it was used to get new insights in scientific and medical applications [71,39,77], to interpret clustering [40], to analyze audio data [75,9], and to compare text classifiers for topic categorization [3].”
Toward Interpretable Machine Learning: Transparent Deep Neural Networks and Beyond (2020, methodology): “As training data we use the recently recorded AudioMNIST [18] dataset, comprised of 30000 audio recordings of spoken digits from 60 different speakers, with 50 repetitions per digit and speaker, in a 48kHz sampling frequency.”
Explaining Deep Neural Networks and Beyond: A Review of Methods and Applications (2021, methodology): “As training data we use the recently recorded AudioMNIST [20] dataset, comprised of 30000 audio recordings of spoken digits from 60 different speakers, with 50 repetitions per digit and speaker, in a 48kHz sampling frequency.”