Search

Backdoor attacks have been recognized as a major AI security threat in deep neural networks (DNNs) recently. The attackers inject backdoors into DNNs during the model training such as federated learning. The infected model behaves normally on the clean samples in AI applications while the backdoors are only activated by the predefined triggers and resulted in the specified results. Most of the existing defensing approaches assume that the trigger settings on different poisoned samples are visible and identical just like a white square in the corner of the image. Besides, the sample-specific triggers are always invisible and difficult to detect in DNNs, which also becomes a great challenge against the existing defensing protocols. In this paper, to address the above problems, we propose a backdoor detecting and mitigating protocol based on a wider separate-then-reunion network (WISERNet) equipped with a cryptographic deep steganalyzer for color images, which detects the backdoors hiding behind the poisoned samples even if the embedding algorithm is unknown and further feeds the poisoned samples into the infected model for backdoor unlearning and mitigation. The experimental results show that our work performs better in the backdoor defensing effect compared to state-of-the-art backdoor defensing methods such as fine-pruning and ABL against three typical backdoor attacks. Our protocol reduces the attack success rate close to 0% on the test data and slightly decreases the classification accuracy on the clean samples within 3%.

In this study, an antidroplet flame retardant system based on FRPET (phosphorus-containing copolyester) is constructed with DOPO-POSS (polyhedral oligomeric silsesquioxane containing DOPO) as an additive flame retardant. It is demonstrated that DOPO-POSS has good dispersibility at a lower amount. When the amount of DOPO-POSS is 9 wt %, the residual char of DOPO-POSS/FRPET at 700 °C increases to 23.56 from 18.16% of FRPET, and the maximum thermal weight loss rate also reduces. What is more is that the limiting oxygen index increases to 33 from 26% of FRPET. The flame burning time is shortened to 4.95 from 20.8 s, the phenomenon of self-extinguishing of the fire occurs, and the vertical combustion level is increased from V-2 to V-0. Compared with FRPET, the peak of the heat release rate decreases by 66.0%, the total heat release decreases by 32.4%, the flame retardancy index (FRI) reaches an excellent value, and the condensed-phase products significantly improve. The Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy-energy dispersive X-ray spectrometry (SEM-EDX), thermogravimetric-FTIR (TG-FTIR), and pyrolysis-gas chromatograph/mass spectrometry (Py-GC/MS) results indicate that DOPO-POSS contributes to the formation of char layers and decomposes to generate free radicals with a quenching effect. In a word, DOPO-POSS is an effective radical trapper and charring agent for PET and exerts a flame retardancy effect in gaseous and condensed phases simultaneously.

The fabrication of bio-based copolyester fiber with adjustable crystallization, orientation structure and mechanical property still remains a great challenge. In this study, a series of copolyester fibers based on terephthalic acid (PTA), ethylene glycol (EG) and l-Lactide (L-LA) were prepared via melt copolymerization and spinning. The resultant PLA-co-PET (PETLA) fibers exhibited tunable structure and property due to the synergistic effects of chemical structure and drawing process. The chemical structure of PETLA was confirmed by NMR, FTIR and XRD, which suggested that the random degree of copolymer increased with LA content and the viscosity decreased with the increase of LA content. The crystallization behavior, melting characteristic, thermal stability and rheological property were investigated by DSC, TGA and rheometer, the results indicated that all the PETLA exhibited the crystallization capacity, melting temperature and thermal stability were slightly affected by LA segment. The synergistic effects of LA segment and spinning process on PETLA structure and property were analyzed by WAXD and SAXS. The breaking strength of PETLA fibers dropped from 5.3 cN/dtex of PET to 2.8 cN/dtex of PET

Improper discharge of slag from mining will pollute the surrounding soil, thereby affecting the ecology and becoming an important global problem. The available copper (ACu) content in polluted soil is an important factor affecting plant growth and development. When investigating a large area of soil with ACu, manual sampling by points and inspection are mainly used, due to the heterogeneity of soil, the efficiency and accuracy are lower. The Unmanned aerial vehicle (UAV) equipped with a hyperspectral sensor as a remote sensing technology is widely used in soil indicator monitoring because of its rapid and convenience. Meanwhile, using the relationship between soil organic matter and available copper has the potential to predict available copper. In this study, we selected the study area with tailings area in the Jianghan Plain of China and used a UAV equipped with a hyperspectral sensor to predict ACu and soil organic matter (SOM) in the soil with two datasets. Firstly, 74 soil samples were collected in the study area, and the ACu and SOM of the soil samples were determined. Second, a hyperspectral image of the study area is obtained using a UAV equipped with a hyperspectral sensor. Thirdly, we combine hyperspectral data with competitive adaptive reweighted sampling (CARS) to obtain feature bands and utilize simulated annealing deep neural network (SA-DNN) to generate estimation models. Finally, maps of the distribution of ACu and SOM in the area were generated using the model. In two datasets, the model of ACu with R

Microplastic and antibiotic contamination are considered an increasing environmental problem in aquatic systems, while little is known about the impact of microplastics and co-pollutant with antibiotics on freshwater vascular plants, particularly the effects of interactions between macrophytes. Here, we performed a mesocosm experiment to evaluate the impact of polyethylene-microplastics and their co-pollutants with ciprofloxacin on the growth and physiological characteristics of Spirodela polyrhiza and Lemna minor and the interactions between these two macrophytes. Our results showed that microplastics alone cannot significantly influence fresh weight and specific leaf area of the two test free-floating macrophytes, but the effects on photosynthetic pigments, malondialdehyde, catalase and soluble sugar contents were species-specific. Ciprofloxacin can significant adverse effects on the growth and physiological traits of the two test macrophytes and microplastic mitigated the toxicity of ciprofloxacin on the two free-floating plants to a certain extent. In addition, our studies showed that microplastics and co-pollutants can influence relative yield and competitiveness of S. polyrhiza and L. minor by directly or indirectly influencing their physiology and growth. Therefore our findings suggest that species-specific sensibility to microplastic and its co-pollutant among free-floating macrophytes may influence macrophyte population dynamics and thereby community structure and ecosystem functioning. And microplastics altered other contaminant behaviours and toxicity, and may directly or indirectly influence macrophytes interactions and community structure. The present study is the first experimental study exploring the effects of microplastics alone and with their co-pollutants on interactions between free-floating macrophytes, which can provide basic theoretical guidance for improving the stability of freshwater ecosystems.

Soil pH and carbonates (CaCO

The dual-function of a wavelength beam splitter and a power beam splitter is desired in both classical optics and quantum optics. We propose a triple-band large-spatial-separation beam splitter at visible wavelengths using a phase-gradient metasurface in both the x- and y-directions. Under x-polarized normal incidence, the blue light is split in the y-direction into two equal-intensity beams owing to the resonance inside a single meta-atom, the green light is split in the x-direction into another two equal-intensity beams owing to the size variation between adjacent meta-atoms, while the red light passes directly without splitting. The size of the meta-atoms was optimized based on their phase response and transmittance. The simulated working efficiencies under normal incidence are 68.1%, 85.0%, and 81.9% at the wavelengths of 420 nm, 530 nm, and 730 nm, respectively. The sensitivities of the oblique incidence and polarization angle are also discussed.

Urban rivers are often narrow, and general remote sensing data cannot meet the needs of water quality monitoring. In the process of monitoring of river water quality by remote sensing, the spectral and spatial dimension of satellite-borne images cannot be taken into consideration at the same time, making fine pollution monitoring of urban rivers difficult. Transparency is one of the core indicators for evaluating water quality, and hyperspectral remote sensing data are rich in spectral information and can be used for quantitative transparency estimation. The application of unmanned aerial vehicles (UAV)remote sensing effectively makes up for the deficiencies in satellite remote sensing monitoring. Aiming at this problem, this paper proposed the use of the eXtreme Gradient Boosting (XGBoost) regression algorithm for the quantitative inversion of urban river transparency. The spatial resolution of the collected imagery is 18.5 cm, which is suitable for urban rivers that are almost ten meters wide. Compared with five traditional empirical models, integrated algorithms such as gradient regression and random forest get much better results. Moreover, the accuracy of transparency estimation using the XGBoost regression algorithm was significantly improved, and the inversion model R2 in both study areas reached over 0.97. Finally, the established transparency inversion models were used to generate transparency distribution maps of the two study areas. The results showed that the distribution of the water transparency was consistent with the results of the field monitoring, indicating that it is feasible to use the XGBoost algorithm for the inversion of urban river transparency in UAV-borne hyperspectral imagery.

The effects of antibiotics on bacterial infections are gradually weakened, leading to the wide development of nanoparticle-based antibacterial agents with unique physical and chemical properties and antibacterial mechanisms different from antibiotics. In this study, we fabricated the uniform and stable graphene oxide (GO)/Ni colloidal nanocrystal cluster (NCNC) nanocomposite by electrostatic self-assembly and investigated its synergistic antibacterial activity against

Deep neural networks (DNNs) have profoundly changed our lifeways in recent years. The cost of training a complicated DNN model is always overwhelming for most users with limited computation and storage resources. Consequently, an increasing number of people are considering to resort to a cloud for an outsourced DNN model training. However, the DNN models training process outsourced to the cloud faces privacy and security issues due to the semi-honest and malicious cloud environments. To preserve the privacy of the data and the parameters in DNN models during the outsourced training and to detect whether the models are injected with backdoors, this paper presents DeepGuard, a framework of privacy-preserving backdoor detection and identification in an outsourced cloud environment for multi-participant computation. In particular, we design a privacy-preserving reverse engineering algorithm for recovering the triggers and detecting the backdoor attacks among three cooperative but non-collusion servers. Moreover, we propose a backdoor identification algorithm adapting to single-label and multi-label attack detection. Finally, extensive experiments on the prevailing datasets such as MNIST, SVHN, and GTSRB confirm the effectiveness and efficiency of backdoor detection and identification in a privacy-preserving DNN model.

Hydrogel dressings are highly biocompatible and can maintain a moist wound environment, suggesting constructing an efficient multi-modal antibacterial hydrogel platform is a promising strategy for treating bacterial wound infections. In this work, a composite Ag