Up-converting rare-earth nanophosphors (UCNPs) have great potential to revolutionize biological luminescent labels, but their use has been limited by difficulties in obtaining UCNPs that are biocompatible. To address this problem, we have developed a simple and versatile strategy for converting hydrophobic UCNPs into water-soluble and carboxylic acid-functionalized analogues by directly oxidizing oleic acid ligands with the Lemieux-von Rudloff reagent. This oxidation process has no obvious adverse effects on the morphologies, phases, compositions and luminescent capabilities of UCNPs. Furthermore, as revealed by Fourier transform infrared (FTIR) and NMR results, oleic acid ligands on the surface of UCNPs can be oxidized into azelaic acids (HOOC(CH2)7COOH), which results in the generation of free carboxylic acid groups on the surface. The presence of free carboxylic acid groups not only confers high solubility in water, but also allows further conjugation with biomolecules such as streptavidin. A highly sensitive DNA sensor based on such streptavidin-coupled UCNPs have been prepared, and the demonstrated results suggest that these biocompatible UCNPs have great superiority as luminescent labeling materials for biological applications.

No abstract

No abstract

Deep Neural Networks (DNNs) are vulnerable to Neural Trojan (NT) attacks where the adversary injects malicious behaviors during DNN training. This type of ‘backdoor’ attack is activated when the input is stamped with the trigger pattern specified by the attacker, resulting in an incorrect prediction of the model. Due to the wide application of DNNs in various critical fields, it is indispensable to inspect whether the pre-trained DNN has been trojaned before employing a model. Our goal in this paper is to address the security concern on unknown DNN to NT attacks and ensure safe model deployment. We propose DeepInspect, the first black-box Trojan detection solution with minimal prior knowledge of the model. DeepInspect learns the probability distribution of potential triggers from the queried model using a conditional generative model, thus retrieves the footprint of backdoor insertion. In addition to NT detection, we show that DeepInspect’s trigger generator enables effective Trojan mitigation by model patching. We corroborate the effectiveness, efficiency, and scalability of DeepInspect against the state-of-the-art NT attacks across various benchmarks. Extensive experiments show that DeepInspect offers superior detection performance and lower runtime overhead than the prior work.

A new homocysteine-selective sensor based on the iridium(III) complex Ir(pba)2(acac) (Hpba = 4-(2-pyridyl)benzaldehyde; acac = acetylacetone) was synthesized, and its' photophysical properties were studied. Upon the addition of homocysteine (Hcy) to a semi-aqueous solution of Ir(pba)2(acac), a color change from orange to yellow and a luminescent variation from deep red to green were evident to the naked eye. The blue-shift of the absorption spectrum and enhancement of the phosphorescence emission upon the addition of Hcy can be attributed to the formation of a thiazinane group by selective reaction of the aldehyde group of Ir(pba)2(acac) with Hcy, which was confirmed by 1H NMR studies. Importantly, Ir(pba)2(acac) shows uniquely luminescent recognition of Hcy over other amino acids (including cysteine) and thiol-related peptides (reduced glutathione), in agreement with the higher luminescent quantum yield of the adduct of Ir(pba)2(acac) with Hcy (0.038) compared with that of the adduct with Cys (~0.002). Both surface charge analysis and the electrochemical measurement indicated that a photoinduced electron-transfer process for Ir(pba)2(acac)-Cys might be responsible for the high specificity of Ir(pba)2(acac) toward Hcy over Cys.