Query-Efficient Black-Box Adversarial Attacks Guided by a Transfer-Based Prior.

Adversarial attacks have been extensively studied in recent years since they can identify the vulnerability of deep learning models before deployed. In this paper, we consider the black-box adversarial setting, where the adversary needs to craft adversarial examples without access to the gradients of a target model. Previous methods attempted to approximate the true gradient either by using the transfer gradient of a surrogate white-box model or based on the feedback of model queries. However, the existing methods inevitably suffer from low attack success rates or poor query efficiency since it is difficult to estimate the gradient in a high-dimensional input space with limited information. To address these problems and improve black-box attacks, we propose two prior-guided random gradient-free (PRGF) algorithms based on biased sampling and gradient averaging, respectively. Our methods can take the advantage of a transfer-based prior given by the gradient of a surrogate model and the query information simultaneously. Through theoretical analyses, the transfer-based prior is appropriately integrated with model queries by an optimal coefficient in each method. Extensive experiments demonstrate that, in comparison with the alternative state-of-the-arts, both of our methods require much fewer queries to attack black-box models with higher success rates.

A solid phase microextraction Arrow with zirconium metal-organic framework/molybdenum disulfide coating coupled with gas chromatography-mass spectrometer for the determination of polycyclic aromatic hydrocarbons in fish samples.

Solid phase microextraction (SPME) Arrows with enlarged sorption phases and Arrow-shaped tips are good alternatives to classic SPME fibers. Developing SPME Arrows with functionalized coatings has become a growing area of SPME research. In this study, a zirconium metal organic framework/molybdenum disulfide (UiO-66/MoS2) composite coating modified SPME Arrow was developed and coupled with gas chromatography-mass spectrometry for the headspace extraction and detection of 16 polycyclic aromatic hydrocarbons (PAHs). The coating preparation can be easily manipulated by the assembly strategy using silicone gels to adhere UiO-66/MoS2 powder onto the surface of the Arrows. The prepared UiO-66/MoS2 coated SPME Arrows exhibited clearly enhanced adsorption capacity (2.1-4.5 folds) and an improved degree of affinity to PAHs species when compared to commercial PDMS/CARS/DVB SPME Arrows or fibers. The improved mechanism may be due to the high specific area, hierarchical micropores and mesopores, and the largely increased and immobilized amount of UiO-66/MoS2 coating on the stainless-steel Arrow. Under optimized conditions, the SPME Arrow-based assay was successfully applied to determine PAH levels in fish samples with satisfactory recoveries of 80.2-101% and relative standard deviations (RSDs) of less than 6%. The detection limits of PAHs were between 0.11-1.40 ng kg-1. The coating showed satisfied reproducibility and repeatability (RSD <8.6%) with only 10 mL of 10% (v/v) acetone in water used as the extraction phase in an effort to be environmentally friendly. All of the results showed that the method is simple, sensitive, and robust. SPME Arrows with UiO-66/MoS2 coatings can provide new opportunities for the efficient extraction of persistent organic pollutants in food safety applications.