Sensitivity-Characterised Activity Neurogram (SCAN) for Visualising and Understanding the Inner Workings of Deep Neural Network

Khe Chai SIM
2016 IEICE transactions on information and systems  
Deep Neural Network (DNN) is a powerful machine learning model that has been successfully applied to a wide range of pattern classification tasks. Due to the great ability of the DNNs in learning complex mapping functions, it has been possible to train and deploy DNNs pretty much as a black box without the need to have an in-depth understanding of the inner workings of the model. However, this often leads to solutions and systems that achieve great performance, but offer very little in terms of
more » ... how and why they work. This paper introduces Sensitivity-characterised Activity Neorogram (SCAN), a novel approach for understanding the inner workings of a DNN by analysing and visualising the sensitivity patterns of the neuron activities. SCAN constructs a lowdimensional visualisation space for the neurons so that the neuron activities can be visualised in a meaningful and interpretable way. The embedding of the neurons within this visualisation space can be used to compare the neurons, both within the same DNN and across different DNNs trained for the same task. This paper will present the observations from using SCAN to analyse DNN acoustic models for automatic speech recognition. key words: deep neural network, visualisation, interpretability Khe Chai Sim received the B.A. & M.Eng. degrees in electrical and information sciences from the University of Cambridge, U.K., in 2001. He then received the M.Phil. degree in computer speech, text and internet technology in 2002 from the same university. He joined the Machine
doi:10.1587/transinf.2016sli0001 fatcat:lwrzobq3wzelre7mbx4wyupccu