Interpreting and Explaining Deep Neural Networks for Classification of Audio Signals

Interpretability of deep neural networks is a recently emerging area of machine learning research targeting a better understanding of how models perform feature selection and derive their classification decisions. This paper explores the interpretability of neural networks in the audio domain by using the previously proposed technique of layer-wise relevance propagation (LRP). We present a novel audio dataset of English spoken digits which we use for classification tasks on spoken digits and speaker's gender. We use LRP to identify relevant features for two neural network architectures that process either waveform or spectrogram representations of the data. Based on the relevance scores obtained from LRP, hypotheses about the neural networks' feature selection are derived and subsequently tested through systematic manipulations of the input data. The results confirm that the networks are highly reliant on features marked as relevant by LRP.

[1]  Biing-Hwang Juang,et al.  Fundamentals of speech recognition , 1993, Prentice Hall signal processing series.

[2]  Sanjeev Khudanpur,et al.  Librispeech: An ASR corpus based on public domain audio books , 2015, 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[3]  Alexander Binder,et al.  Evaluating the Visualization of What a Deep Neural Network Has Learned , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[4]  M. A. Anusuya,et al.  Speech Recognition by Machine, A Review , 2010, ArXiv.

[5]  Jae S. Lim,et al.  Signal estimation from modified short-time Fourier transform , 1983, ICASSP.

[6]  D.R. Reddy,et al.  Speech recognition by machine: A review , 1976, Proceedings of the IEEE.

[7]  Anders Eriksson,et al.  The frequency range of the voice fundamental in the speech of male and female adults , 1993 .

[8]  K. Nagata.  Spoken digit recognizer for Japanese language. , 1963 .

[9]  Andrea Vedaldi,et al.  Interpretable Explanations of Black Boxes by Meaningful Perturbation , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[10]  Tara N. Sainath,et al.  Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups , 2012, IEEE Signal Processing Magazine.

[11]  Alexander Binder,et al.  Analyzing Classifiers: Fisher Vectors and Deep Neural Networks , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  Alexander Binder,et al.  Unmasking Clever Hans predictors and assessing what machines really learn , 2019, Nature Communications.

[13]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[14]  Motoaki Kawanabe,et al.  How to Explain Individual Classification Decisions , 2009, J. Mach. Learn. Res..

[15]  Trevor Darrell,et al.  Caffe: Convolutional Architecture for Fast Feature Embedding , 2014, ACM Multimedia.

[16]  Klaus-Robert Müller,et al.  Interpretable deep neural networks for single-trial EEG classification , 2016, Journal of Neuroscience Methods.

[17]  Klaus-Robert Müller,et al.  Explaining Recurrent Neural Network Predictions in Sentiment Analysis , 2017, WASSA@EMNLP.

[18]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[19]  Pascal Vincent,et al.  Visualizing Higher-Layer Features of a Deep Network , 2009 .

[20]  Yann LeCun,et al.  The mnist database of handwritten digits , 2005 .

[21]  Ruben Gonzalez.  Better Than MFCC Audio Classification Features , 2013 .

[22]  Honglak Lee,et al.  Unsupervised feature learning for audio classification using convolutional deep belief networks , 2009, NIPS.

[23]  Brian Kingsbury,et al.  New types of deep neural network learning for speech recognition and related applications: an overview , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[24]  Alexander Binder,et al.  Explaining nonlinear classification decisions with deep Taylor decomposition , 2015, Pattern Recognit..

[25]  Aren Jansen,et al.  CNN architectures for large-scale audio classification , 2016, 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[26]  Nacereddine Hammami,et al.  Tree distribution classifier for automatic spoken Arabic digit recognition , 2009, 2009 International Conference for Internet Technology and Secured Transactions, (ICITST).

[27]  John J. Godfrey,et al.  SWITCHBOARD: telephone speech corpus for research and development , 1992, [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[28]  Been Kim,et al.  Towards A Rigorous Science of Interpretable Machine Learning , 2017, 1702.08608.

[29]  Wojciech Samek,et al.  Methods for interpreting and understanding deep neural networks , 2017, Digit. Signal Process..

[30]  Johannes Gehrke,et al.  Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-day Readmission , 2015, KDD.

[31]  Jonathan G. Fiscus,et al.  DARPA TIMIT:: acoustic-phonetic continuous speech corpus CD-ROM, NIST speech disc 1-1.1 , 1993 .

[32]  Alexandre Tkatchenko,et al.  Quantum-chemical insights from deep tensor neural networks , 2016, Nature Communications.

[33]  Yee Whye Teh,et al.  Unsupervised Discovery of Nonlinear Structure Using Contrastive Backpropagation , 2006, Cogn. Sci..

[34]  Wei Dai,et al.  Very deep convolutional neural networks for raw waveforms , 2016, 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[35]  Xinlei Chen,et al.  Visualizing and Understanding Neural Models in NLP , 2015, NAACL.

[36]  Alexander Binder,et al.  On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation , 2015, PloS one.

[37]  Anna Shcherbina,et al.  Not Just a Black Box: Learning Important Features Through Propagating Activation Differences , 2016, ArXiv.