From hate to harmony: Leveraging large language models for safer speech in times of COVID-19 crisis.

This study investigates the rampant spread of offensive and derogatory language during the COVID-19 pandemic and aims to mitigate it through machine learning. Employing advanced Large Language Models (LLMs), the research develops a sophisticated framework adept at detecting and transforming abusive and hateful speech. The project begins by meticulously compiling a dataset, focusing specifically on Chinese language abuse and hate speech. It incorporates an extensive list of 30 pandemic-related terms, significantly enriching the resources available for this type of research. A two-tier detection model is then introduced, achieving a remarkable accuracy of 94.42 % in its first phase and an impressive 81.48 % in the second. Furthermore, the study enhances paraphrasing efficiency by integrating generative AI techniques, primarily Large Language Models, with a Latent Dirichlet Allocation (LDA) topic model. This combination allows for a thorough analysis of language before and after modification. The results highlight the transformative power of these methods. They show that the rephrased statements not only reduce the initial hostility but also preserve the essential themes and meanings. This breakthrough offers users effective rephrasing suggestions to prevent the spread of hate speech, contributing to more positive and constructive public discourse.

Nitric oxide releasing poly(vinyl alcohol)/S-nitrosated keratin film as a potential vascular graft.

Nitric oxide (NO) releasing vascular graft is promising due to its merits of thromboembolism reduction and endothelialization promotion. In this study, keratin-based NO donor of S-nitrosated keratin (KSNO) was blended with poly(vinyl alcohol) (PVA) and further crosslinked with sodium trimetaphosphate (STMP) to afford PVA/KSNO biocomposite films. These films could release NO sustainably for up to 10 days, resulting in the promotion of HUVECs growth and the inhibition of HUASMCs growth. In addition, these films displayed good blood compatibility and antibacterial activity. Taken together, these films have potential applications in vascular grafts.

Iron Competition as an Important Mechanism of Pulcherrimin-Producing Metschnikowia sp. Strains for Controlling Postharvest Fungal Decays on Citrus Fruit.

This study identified and tested fruit-isolated Metschnikowia yeasts against three major postharvest citrus pathogens, namely, Penicillium digitatum, Penicillium italicum, and Geotrichum citri-aurantii, and further evaluated the impact of FeCl3 on the biocontrol efficiency of pulcherrimin-producing M. pulcherrima strains. Based on the characterization of the pigmented halo surrounding the colonies and the analysis of the D1/D2 domain of 26S rDNA, a total of 46 Metschnikowia sp. were screened and identified. All 46 Metschnikowia strains significantly inhibited the hyphal growth of Penicillium digitatum, Penicillium italicum, and Geotrichum citri-aurantii, and effectively controlled the development of green mold, blue mold and sour rot of citrus fruit. The introduction of exogenous FeCl3 at certain concentrations did not significantly impact the pulcherriminic acid (PA) production of pigmented M. pulcherrima strains, but notably diminished the size of pigmented zones and the biocontrol efficacy against the three pathogens. Iron deficiency sensitivity experiments revealed that P. digitatum and P. italicum exhibited higher sensitivity compared to G. citri-aurantii, indicating that iron dependence varied among the three pathogens. These results suggested that M. pulcherrima strains, capable of producing high yields of PA, possessed great potential for use as biocontrol agents against postharvest citrus diseases. The biocontrol efficacy of these yeasts is mainly attributed to their ability to competitively deplete iron ions in a shared environment, with the magnitude of their pigmented halo directly correlating to their antagonistic capability. It is worth noting that the level of sensitivity of pathogens to iron deficiency might also affect the biocontrol effect of pulcherrimin-producing M. pulcherrima.

Effect of alkaline hydrogen peroxide assisted with two modification methods on the physicochemical, structural and functional properties of bagasse insoluble dietary fiber.

Bagasse is one of major by-product of sugar mills, but its utilization is limited by the high concentration of lignin. In this study, the optimal alkaline hydrogen peroxide (AHP) treatment conditions were determined by the response surface optimization method. The results showed that the lignin removal rate was 62.23% and the solid recovery rate was 53.76% when bagasse was prepared under optimal conditions (1.2% H2O2, 0.9% NaOH, and 46°C for 12.3 h), while higher purity of bagasse insoluble dietary fiber (BIDF) was obtained. To further investigate the modification effect, AHP assisted with high-temperature-pressure cooking (A-H) and enzymatic hydrolysis (A-E) were used to modify bagasse, respectively. The results showed that the water holding capacity (WHC), oil holding capacity (OHC), bile salt adsorption capacity (BSAC), and nitrite ion adsorption capacity (NIAC) were significantly improved after A-H treatment. With the A-E treatment, cation exchange capacity (CEC) and BSAC were significantly increased, while WHC, OHC, and glucose adsorption capacity (GAC) were decreased. Especially, the highest WHC, OHC, BSAC and NIAC were gained by A-H treatment compared to the A-E treatment. These changes in the physicochemical and functional properties of bagasse fiber were in agreement with the microscopic surface wrinkles and pore structure, crystallinity and functional groups. In summary, the A-H modification can effectively improve the functional properties of bagasse fiber, which potentially can be applied further in the food industry.

Facile fabrication of copper-incorporating poly(ε-caprolactone)/keratin mats for tissue-engineered vascular grafts with the potential of catalytic nitric oxide generation.

Tissue-engineered vascular grafts (TEVGs) provide a new alternative for vascular construction. Nitric oxide (NO) is capable of promoting vascular tissue regeneration and reducing restenosis caused by vascular implantation. Therefore, in situ production of NO by catalytic decomposition of the endogenous donor is a promising strategy to fabricate a TEVG. In this study, poly(ε-caprolactone) (PCL) was first electrospun with keratin (Ker) to afford PCL/Ker mats and then incorporated with Cu(II) ions through multiple interactions. This strategy is very simple, green, and facile. Particularly, the incorporated Cu(II) ions were partially reduced to Cu(I) ions due to the reducibility of keratin. The chelated copper ions were expected to catalyze the generation of NO from endogenous S-nitrosothiol (RSNO). As a result, PCL/Ker-Cu mats selectively accelerated the adhesion, migration, and growth of human umbilical vein endothelial cells (HUVECs), while inhibiting the proliferation of human umbilical artery smooth muscle cells (HUASMCs). Furthermore, these mats exhibited excellent blood compatibility and significant antibacterial activity. Vascular implantation in vivo indicated that the tubular mats could inhibit thrombus formation and retain patency for 3 months after implantation in the rabbit carotid artery. More importantly, vascular remodeling was observed during follow-up, including a complete endothelium and smooth muscle layer. Taken together, the PCL/Ker-Cu mats have great potential application in vascular tissue regeneration.

Biocompatible and photocrosslinkable poly(ethylene glycol)/keratin biocomposite hydrogels.

A biocompatible hydrogel is ideal for tissue engineering and regeneration. In this study, methacrylated keratin (KerMA) was synthesized for the first time and then blended with poly(ethylene glycol dimethacrylate) (PEGDMA) to form hydrogel through photocrosslinking. The chemical structure, gelation time, swelling behavior, hydrophilicity, cytotoxicity, and 3D printability of PEGDMA/KerMA hydrogels were characterized and exploited. The PEGDMA/KerMA hydrogels performed good cytocompatibility, providing potential applications for tissue engineering and regenerative medicine.

Molecular spectra of a D-π-A typed polydentate ligand chromophore and its simultaneous response to trace Cu2+ and Co2.

A D-π-A conjugated polydentate ligand chromophore, N-8'-quinolyl-2,4,6- trihydroxyl benzamide (NQTB), was identified and synthesized using tri-hydroxyl phenol as donated-electron group, N-heterocycle quinoline as accepted-electron one and CN bond as bridged one. It was expected to chelate some heavy metal ions with prominent colorimetric or spectral changes. After its UV-vis absorption spectrum was investigated in detail, it was noted that NQTB possessed excellent spectral recognition ability to Cu2+ and Co2+ from other coexisting ions in aqueous. Under the optimized conditions, NQTB could simultaneously discriminate trace Cu2+ and Co2+ in environmental aqueous samples with low detection limits (1.9 × 10-8 mol/L and 5.7 × 10-8 mol/L) and satisfying analytical precisions (R.S.D. ≤3.3% and ≤2.6%) respectively. The sensing mechanism was confirmed to form some stable 5-membered-co-6-membered condensed rings between Cu2+/Co2+ and O/N atoms in NQTB.

To Explore Intracerebral Hematoma with a Hybrid Approach and Combination of Discriminative Factors.

OBJECTIVES: To find discriminative combination of influential factors of Intracerebral hematoma (ICH) to cluster ICH patients with similar features to explore relationship among influential factors and 30-day mortality of ICH. METHODS: The data of ICH patients are collected. We use a decision tree to find discriminative combination of the influential factors. We cluster ICH patients with similar features using Fuzzy C-means algorithm (FCM) to construct a support vector machine (SVM) for each cluster to build a multi-SVM classifier. Finally, we designate each testing data into its appropriate cluster and apply the corresponding SVM classifier of the cluster to explore the relationship among impact factors and 30-day mortality. RESULTS: The two influential factors chosen to split the decision tree are Glasgow coma scale (GCS) score and Hematoma size. FCM algorithm finds three centroids, one for high danger group, one for middle danger group, and the other for low danger group. The proposed approach outperforms benchmark experiments without FCM algorithm to cluster training data. CONCLUSIONS: It is appropriate to construct a classifier for each cluster with similar features. The combination of factors with significant discrimination as input variables should outperform that with only single discriminative factor as input variable.