Search for: All records

critical to reveal a blackbox model’s decision-making process from raw data to prediction. In this article, we use two real datasets, the MNIST handwritten digits and MIT-BIH Electrocardiogram (ECG) signals, to motivate key characteristics of discriminative features, namely adaptiveness, predictive importance and effectiveness. Then, we develop a localization framework based on adversarial attacks to effectively localize discriminative features. In contrast to existing heuristic methods, we also provide a statistically guaranteed interpretability of the localized features by measuring a generalized partial R2. We apply the proposed method to the MNIST dataset and the MIT-BIH dataset with a convolutional auto-encoder. In the first, the compact image regions localized by the proposed method are visually appealing. Similarly, in the second, the identified ECG features are biologically plausible and consistent with cardiac electrophysiological principles while locating subtle anomalies in a QRS complex that may not be discernible by the naked eye. Overall, the proposed method compares favorably with state-of-the-art competitors. Accompanying this paper is a Python library dnn-locate that implements the proposed approach. 

Statistical inference of directed relations given some unspecified interventions (i.e., the intervention targets are unknown) is challenging. In this article, we test hypothesized directed relations with unspecified interventions. First, we derive conditions to yield an identifiable model. Unlike classical inference, testing directed relations requires identifying the ancestors and relevant interventions of hypothesis-specific primary variables. To this end, we propose a peeling algorithm based on nodewise regressions to establish a topological order of primary variables. Moreover, we prove that the peeling algorithm yields a consistent estimator in low-order polynomial time. Second, we propose a likelihood ratio test integrated with a data perturbation scheme to account for the uncertainty of identifying ancestors and interventions. Also, we show that the distribution of a data perturbation test statistic converges to the target distribution. Numerical examples demonstrate the utility and effectiveness of the proposed methods, including an application to infer gene regulatory networks. The R implementation is available at https://github.com/chunlinli/intdag. 

There is no doubt that there is an increase in the penetration of electrical energy into the operation of high-speed railway systems (HSR). This is even more pronounced with the increasing trends in smart electric multiple units (EMU). The operational speed serves as a metric for punctuality and safety, as well as a critical element to maintain the balance between energy supply and consumption. The speed-based regenerative energy from EMU’s braking mode could be utilized in the restoration of system operation in the aftermath of a failure. This paper optimizes the system resiliency with respect to the operational speed for the purpose of restoration by minimizing the total cost of implementing recovery measures. By simultaneously valuating the dual-impact of any given fault on the speed deterioration level from the railway operation systems (ROS) side and the power supply and demand unbalance level from the railway power systems (RPS) side, this process develops an adaptive two-dimension risk assessment scheme for prioritizing the handling of different operational zones that are cascaded in the system. With the aid of an integrated speed-based resilience cost model, we determine the optimal resilience time, speed modification plan, and energy allocation strategy. The outcome from implementing this routine in a real-world HSR offers a pioneering decision-making strategy and perspective on optimizing the resilience of an integrated system. 

Intraparticle charge delocalization occurs when metal nanoparticles are functionalized with organic capping ligands through conjugated metal-ligand interfacial bonds. In this study, metal nanoparticles of 5d metals (Ir, Pt, and Au) and 4d metals (Ru, Rh, and Pd) were prepared and capped with ethynylphenylacetylene and the impacts of the number of metal d electrons on the nanoparticle optoelectronic properties were examined. Both FTIR and photoluminescence measurements indicate that intraparticle charge delocalization was enhanced with the increase of the number of d electrons in the same period with palladium being an exception.