Deep learning (DL) has achieved remarkable progress over the past decade and been widely applied to many safety-critical applications. However, the robustness of DL systems recently receives great concerns, such as adversarial examples against computer vision systems, which could potentially result in severe consequences. Adopting testing techniques could help to evaluate the robustness of a DL system and therefore detect vulnerabilities at an early stage. The main challenge of testing such systems is that its runtime state space is too large: if we view each neuron as a runtime state for DL, then a DL system often contains massive states, rendering testing each state almost impossible. For traditional software, combinatorial testing (CT) is an effective testing technique to reduce the testing space while obtaining relatively high defect detection abilities. In this paper, we perform an exploratory study of CT on DL systems. We adapt the concept in CT and propose a set of coverage criteria for DL systems, as well as a CT coverage guided test generation technique. Our evaluation demonstrates that CT provides a promising avenue for testing DL systems. We further pose several open questions and interesting directions for combinatorial testing of DL systems.
Cardiac arrhythmia is associated with abnormal electrical activities of the heart, which can be reflected by altered characteristics of electrocardiogram (ECG). Due to the simplicity and non-invasive nature, the ECG has been widely used for detecting arrhythmias and there is an urgent need for automatic ECG detection. Up to date, some algorithms have been proposed for automatic classification of cardiac arrhythmias based on the features of the ECG; however, their stratification rate is still poor due to unreliable features of signal characteristics or limited generalization capability of the classifier, and therefore, it remains a challenge for automatic diagnosis of arrhythmias. In this paper, we propose a new method for automatic classification of arrhythmias based on deep neural networks (DNNs). The two DNN models constitutive of residual convolutional modules and bidirectional long short-term memory (LSTM) layers are trained to extract features from raw ECG signals. The extracted features are concatenated to form a feature vector which is trained to do the final classification. The algorithm is evaluated based on the test set of China Physiological Signal Challenge (CPSC) dataset with F 1 measure regarded as the harmonic mean between the precision and recall. The resulting overall F 1 score is 0.806, F AF score is 0.914 for atrial fibrillation (AF), F Block score is 0.879 for block, F PC and F ST scores are 0.801 and 0.742 for premature contraction and ST-segment change, which demonstrates a good performance that may have potential practical applications.
INDEX TERMSCardiac arrhythmia, electrocardiogram (ECG), deep neural networks (DNNs), deep residual network, bidirectional long short-term memory (LSTM).
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.