Fuzzy Self-Adaptation of Mission-Critical Software Under Uncertainty

Mission-critical software （MCS） must provide continuous, online services to ensure the successful accomplish- ment of critical missions. Self-adaptation is particularly desirable for assuring the quality of service （QoS） and availability of MCS under uncertainty. Few techniques have insofar addressed the issue of MCS self-adaptation, and most existing approaches to software self-adaptation fail to take into account uncertainty in the self-adaptation loop. To tackle this problem, we propose a fuzzy control based approach, i.e., Software Fuzzy Self-Adaptation （SFSA）, with a view to deal with the challenge of MCS self-adaptation under uncertainty. First, we present the SFSA conceptual framework, consisting of sensing, deciding and acting stages, and establish the formal model of SFSA to lay a rigorous and mathematical foundation of our approach. Second, we develop a novel SFSA implementation technology as well as its supporting tool, i.e., the SFSA toolkit, to automate the realization process of SFSA. Finally, we demonstrate the effectiveness of our approach through the development of an adaptive MCS application in process control systems. Validation experiments show that the fuzzy control based approach proposed in this work is effective and with low overheads.

Predicted Robustness as QoS for Deep Neural Network Models

The adoption of deep neural network(DNN)model as the integral part of real-world software systems necessitates explicit consideration of their quality-of-service(QoS).It is well-known that DNN models are prone to adversarial attacks,and thus it is vitally important to be aware of how robust a modeFs prediction is for a given input instance.A fragile prediction,even with high confidence,is not trustworthy in light of the possibility of adversarial attacks.We propose that DNN models should produce a robustness value as an additional QoS indicator,along with the confidence value,for each prediction they make.Existing approaches for robustness computation are based on adversarial searching,which are usually too expensive to be excised in real time.In this paper,we propose to predict,rather than to compute,the robustness measure for each input instance.Specifically,our approach inspects the output of the neurons of the target model and trains another DNN model to predict the robustness.We focus on convolutional neural network(CNN)models in the current research.Experiments show that our approach is accurate,with only 10%-34%additional errors compared with the offline heavy-weight robustness analysis.It also significantly outperforms some alternative methods.We further validate the effectiveness of the approach when it is applied to detect adversarial attacks and out-of-distribution input.Our approach demonstrates a better performance than,or at least is comparable to,the state-of-the-art techniques.

Poly(vinylidene fluoride) membrane based thin film microextraction for enrichment of benzoylurea insecticides from water samples followed by their determination with HPLC

Thin-film microextraction（TFME）,a new geometry for solid-phase microextraction,has become an attractive sample-preparation technique.Compared to other microextraction approaches,the sensitivity of this technique was enhanced without sacrificing the sampling time due to the high surface area-tovolume ratio together with the increase of extraction-phase volume.In this paper,a new TFME method based on poly（vinylidene fluoride） membrane was developed for the extraction of benzoylurea insecticides（diflubenzuron,triflumuron,hexaflumuron and teflubenzuron） from water samples followed by their determination with high performance liquid chromatography-diode array detection.Under the optimal conditions,good linearity was observed over the concentration range of 0.5-100.0 ng/mL with correlation coefficient greater than 0.9994.The limits of detection（S/N = 3） of the method for the target analytes were 0.1 ng/mL.Mean recoveries ranged from 87.7% to 103.9% with relative standard deviations lower than 6.5%.The results indicated that the developed TFME method is simple,efficient,and cost effective.