Word distance assisted dual graph convolutional networks for accurate and fast aspect-level sentiment analysis.

Aspect-level sentiment analysis can provide a fine-grain sentiment classification for inferring the sentiment polarity of specific aspects. Graph convolutional network (GCN) becomes increasingly popular because its graph structure can characterize the words' correlation for extracting more sentiment information. However, the word distance is often ignored and cause the cross-misclassification of different aspects. To address the problem, we propose a novel dual GCN structure to take advantage of word distance, syntactic information, and sentiment knowledge in a joint way. The word distance is not only used to enhance the syntactic dependency tree, but also to construct a new graph with semantic knowledge. Then, the two kinds of word distance assisted graphs are fed into two GCNs for further classification. The comprehensive results on two self-collected Chinese datasets (MOOC comments and Douban book reviews) as well as five open-source English datasets, demonstrate that our proposed approach achieves higher classification accuracy than the state-of-the-art methods with up to 1.81x training acceleration.

dm-GAN: Distributed multi-latent code inversion enhanced GAN for fast and accurate breast X-ray image automatic generation.

Breast cancer seriously threatens women's physical and mental health. Mammography is one of the most effective methods for breast cancer diagnosis via artificial intelligence algorithms to identify diverse breast masses. The popular intelligent diagnosis methods require a large amount of breast images for training. However, collecting and labeling many breast images manually is extremely time consuming and inefficient. In this paper, we propose a distributed multi-latent code inversion enhanced Generative Adversarial Network (dm-GAN) for fast, accurate and automatic breast image generation. The proposed dm-GAN takes advantage of the generator and discriminator of the GAN framework to achieve automatic image generation. The new generator in dm-GAN adopts a multi-latent code inverse mapping method to simplify the data fitting process of GAN generation and improve the accuracy of image generation, while a multi-discriminator structure is used to enhance the discrimination accuracy. The experimental results show that the proposed dm-GAN can automatically generate breast images with higher accuracy, up to a higher 1.84 dB Peak Signal-to-Noise Ratio (PSNR) and lower 5.61% Fréchet Inception Distance (FID), as well as 1.38x faster generation than the state-of-the-art.

An adversarially consensus model of augmented unlabeled data for cardiac image segmentation (CAU+).

High quality medical images play an important role in intelligent medical analyses. However, the difficulty of acquiring medical images with professional annotation makes the required medical image datasets, very expensive and time-consuming. In this paper, we propose a semi-supervised method, $ {\mathrm{C}\mathrm{A}\mathrm{U}}^{+} $, which is a consensus model of augmented unlabeled data for cardiac image segmentation. First, the whole is divided into two parts: the segmentation network and the discriminator network. The segmentation network is based on the teacher student model. A labeled image is sent to the student model, while an unlabeled image is processed by CTAugment. The strongly augmented samples are sent to the student model and the weakly augmented samples are sent to the teacher model. Second, $ {\mathrm{C}\mathrm{A}\mathrm{U}}^{+} $ adopts a hybrid loss function, which mixes the supervised loss for labeled data with the unsupervised loss for unlabeled data. Third, an adversarial learning is introduced to facilitate the semi-supervised learning of unlabeled images by using the confidence map generated by the discriminator as a supervised signal. After evaluating on an automated cardiac diagnosis challenge (ACDC), our proposed method $ {\mathrm{C}\mathrm{A}\mathrm{U}}^{+} $ has good effectiveness and generality and $ {\mathrm{C}\mathrm{A}\mathrm{U}}^{+} $ is confirmed to have a improves dice coefficient (DSC) by up to 18.01, Jaccard coefficient (JC) by up to 16.72, relative absolute volume difference (RAVD) by up to 0.8, average surface distance (ASD) and 95% Hausdorff distance ($ {HD}_{95} $) reduced by over 50% than the latest semi-supervised learning methods.

Insights into C-C Bond Cleavage Mechanisms in Dichloroacetonitrile Formation during Chlorination of Long-Chain Primary Amines, Amino Acids, and Dipeptides.

Dichloroacetonitrile (DCAN) as one of the potentially prioritized regulated DBPs has drawn great attention; however, understanding its formation, especially the C-C bond cleavage mechanisms, is limited. In this study, DCAN formation mechanisms from long-chain primary amines, amino acids, and dipeptides during chlorination were investigated by a combined computational and experimental approach. The results indicate that nitriles initially generate for all of the above precursors, then they undergo ß-C-hydroxylation or/and α-C-chlorination processes, and finally, DCAN is produced through the Cα-Cß bond cleavage. For the first time, the underlying mechanism of the C-C bond cleavage was unraveled to be electron transfer from the O- anion into its attached C atom in the chlorinated nitriles, leading to the strongly polarized Cα-Cß bond heterocleavage and DCAN- formation. Moreover, DCAN molar yields of precursors studied in the present work were found to be determined by their groups at the γ-site of the amino group, where the carbonyl group including -CO2-, -COR, and -CONHR, the aromatic group, and the -OH group can all dramatically facilitate DCAN formation by skipping over or promoting the time-consuming ß-C-hydroxylation process and featuring relatively lower activation free energies in the C-C bond cleavage. Importantly, 4-amino-2-hydroxybutyric acid was revealed to possess the highest DCAN yield among all the known aliphatic long-chain precursors to date during chlorination. Additionally, enonitriles, (chloro-)isocyanates, and nitriles can be generated during DCAN formation and should be of concern due to their high toxicities.

Trichloramine and Hydroxyl Radical Contributions to Dichloroacetonitrile Formation Following Breakpoint Chlorination.

Breakpoint chlorination is applied to remove ammonia in water treatment. Trichloramine (NCl3) and transient reactive species can be present, but how they affect the formation of nitrogenous disinfection byproducts is unknown. In this study, the dichloroacetonitrile (DCAN) formation mechanisms and pathways involved during breakpoint chlorination (i.e., free chlorine to ammonia molar ratio ≥2.0) were investigated. DCAN formation during breakpoint chlorination of natural organic matter (NOM) isolates was 14.3-20.3 µg/L, which was 2-10 times that in chlorination without ammonia at similar free chlorine residual conditions (2.1-2.9 mg/L as Cl2). The probe tests and electron paramagnetic resonance spectra supported the presence of •OH, •NO, and NCl3 besides free chlorine in breakpoint chlorination. 15N-labeled ammonium-N tests indicated the incorporation of ammonium-N in DCAN formation though ammonia was eliminated during breakpoint chlorination. Aromatic non-nitrogenous moieties, such as phenols (i.e., none DCAN precursors in the free-chlorine-only system), became DCAN precursors during breakpoint chlorination. The reactions involved in reactive nitrogen species, such as •NO/•NO2 and NCl3, led to additional nitrogen sources in DCAN formation, accounting for 36-84% of total nitrogen sources in DCAN formation from NOM isolates and real water samples. Scavenging •OH by tert-butanol reduced DCAN formation by 40-56%, indicating an important role of •OH in transforming DCAN precursors. This study improves the understanding of breakpoint chlorination chemistry.

Rhein and polydimethylsiloxane functionalized carbon/carbon composites as prosthetic implants for bone repair applications.

A major issue in bone tissue engineering is the selection of biocompatible materials for implants, to reduce unwanted inflammatory reactions and promote cell adhesion. Bone tissue growth on suitable biomedical implants can shorten recovery and hospitalization after surgery. Therefore, a method to improve tissue-implant integration and healing would be of scientific and clinical interest. In this work, we permeated polydimethylsiloxane (PDMS) into carbon/carbon (C/C) composites (PDMS-C/C) and then coated it with 4,5-dihydroxyanthraquinone-2-carboxylic acid (rhein) to create rhein-PDMS-C/C to increase its biocompatibility and reduce the occurrence of inflammatory reactions. We measured in vitro adhesion and proliferation of MC3T3-E1 cells and bacteria to evaluate the biocompatibility and antimicrobial properties of C/C, PDMS-C/C, and rhein-PDMS-C/C. In vivo, x-ray and micro-CT evaluation three, six and nine weeks after surgery revealed that rhein-PDMS-C/C was more effective than PDMS-C/C and C/C composite in terms of antibacterial activity, cell adhesion and tissue growth. Compared with C/C and PDMS-C/C, rhein-PDMS-C/C could be suitable for clinical applications for bone tissue engineering.

PEGylation corannulene enhances response of stress through promoting neurogenesis.

Carbon nanomaterials have been used to treat neurodegenerative diseases and neural disorders due to their diverse molecular structures. Corannulene is a three-dimensional π-bowl carbon nanomaterial that is different from planar PAHs, fullerenes and carbon nanotubes, but little is known about its biological functions. Herein, corannulene was functionalized with mPEG-DESP to prepare PEGylation corannulene nanoparticles (PEGylation CoNps). The synthesized PEGylation CoNps shows enhanced solubility and reduced aggregation when compared corannulene. Then, in vivo experiments were performed to determine the effects of PEGylation CoNps on the neural system. We found that PEGylation CoNps treatment increased short resting bouts, decreased locomotion activities and enhanced the response to stress. Most of these behavioral changes suggest that PEGylation CoNps lead to a greater reflection to stress, which is associated with neurotransmitter expression and neurogenesis. In line with the hypothesis, we found that PEGylation CoNps administration enhanced TH, DCX and MAP-2 expression in the hippocampus. These results indicated that PEGylation CoNps enhanced the neurogenesis of mice. Furthermore, pathological analysis showed that PEGylation CoNps caused little inflammation. These findings suggest that PEGylation CoNps are a potential functionalized carbon nanomaterial for promoting neurogenesis.

Antibacterial and anticancer PDMS surface for mammalian cell growth using the Chinese herb extract paeonol(4-methoxy-2-hydroxyacetophenone).

Polydimethylsiloxane (PDMS) is widely used as a cell culture platform to produce micro- and nano-technology based microdevices. However, the native PDMS surface is not suitable for cell adhesion and is always subject to bacterial pollution and cancer cell invasion. Coating the PDMS surface with antibacterial or anticancer materials often causes considerable harm to the non-cancer mammalian cells on it. We have developed a method to fabricate a biocompatible PDMS surface which not only promotes non-cancer mammalian cell growth but also has antibacterial and anticancer activities, by coating the PDMS surface with a Chinese herb extract, paeonol. Coating changes the wettability and the elemental composition of the PDMS surface. Molecular dynamic simulation indicates that the absorption of paeonol onto the PDMS surface is an energy favourable process. The paeonol-coated PDMS surface exhibits good antibacterial activity against both Gram-positive and Gram-negative bacteria. Moreover considerable antibacterial activity is maintained after the coated surface is rinsed or incubated in water. The coated PDMS surface inhibits bacterial growth on the contact surface and promotes non-cancer mammalian cell growth with low cell toxicity; meanwhile the growth of cancer cells is significantly inhibited. Our study will potentially guide PDMS surface modification approaches to produce biomedical devices.

Biostable L-DNA-Templated Aptamer-Silver Nanoclusters for Cell-Type-Specific Imaging at Physiological Temperature.

The high susceptibility of the natural D-conformation of DNA (D-DNA) to nucleases greatly limits the application of DNA-templated silver nanoclusters (Ag NCs) in biological matrixes. Here we demonstrate that the L-conformation of DNA (L-DNA), the enantiomer of D-DNA, can also be used for the preparation of aptamer-Ag NCs. The extraordinary resistance of L-DNA to nuclease digestion confers much higher biostability to these NCs than those templated by D-DNA, thus making cell-type-specific imaging possible at physiological temperatures, using at least 100-times lower Ag NC concentration than reported D-DNA-templated ones. The L-DNA-templated metal NC probes with enhanced biostability might promote the applications of metal nanocluster probes in complex biological systems.