PLoS ONE最新文献

Zero-shot image classification enables the recognition of new categories without requiring additional training data, thereby enhancing the model's generalization capability when specific training are unavailable. This paper introduces a zero-shot image classification framework to recognize new categories that are unseen during training by distilling knowledge from foundation models. Specifically, we first employ ChatGPT and DALL-E to synthesize reference images of unseen categories from text prompts. Then, the test image is aligned with text and reference images using CLIP and DINO to calculate the logits. Finally, the predicted logits are aggregated according to their confidence to produce the final prediction. Experiments are conducted on multiple datasets, including MNIST, SVHN, CIFAR-10, CIFAR-100, and TinyImageNet. The results demonstrate that our method can significantly improve classification accuracy compared to previous approaches, achieving AUROC scores of over 96% across all test datasets. Our code is available at https://github.com/1134112149/MICW-ZIC.

Purpose: To examine trends and potential disparities in North Carolina (NC) driving while impaired by alcohol (alcohol-DWI) license suspensions from 2007-2016. Specific objectives included: 1) examining personal (e.g., race/ethnicity) and contextual (e.g., residential segregation) characteristics of alcohol-DWI license suspensions by suspension duration; and 2) examining trends in annual suspension rates by race/ethnicity, sex, and duration.

Methods: We linked NC administrative licensing and county-level survey data from several sources from 2007-2016. Suspensions were categorized by duration: 1 to <4 years and 4 years or longer (proxies for initial and repeat suspensions, respectively). We calculated counts, percentages, and suspensions rates (per 1,000 person-years) with 95% confidence intervals, examined trends in annual suspension rates by race/ethnicity, sex, and suspension duration.

Results: We identified 220,471 initial and 41,526 repeat license suspensions. Rates among males were three times that of females. 21-24-year-old (rates: 6.9 per 1,000 person-years for initial; 1.5 for repeat) and Black (4.1 for initial; 1.0 for repeat) individuals had the highest suspension rates. We observed decreases in annual initial and repeat suspension rates among males, but only in repeat suspensions for females during the study period. A substantial decrease in annual initial suspension rates was observed among Hispanic individuals relative to other racial/ethnic groups, while annual repeat suspension rates exhibited large decreases for most racial/ethnic groups. The highest overall suspension rates occurred in counties with higher proportions of the population without health insurance and with the highest levels of Black/White residential segregation.

Conclusions: Potential disparities by race/ethnicity and sex existed by alcohol-DWI license suspension duration (i.e., initial vs. repeat suspensions) in NC. Contextual characteristics associated with suspensions, including a high degree of residential segregation, may provide indications of underlying structures and mechanisms driving potential disparities in alcohol-DWI outcomes.

Background: Nowadays, the contradiction between the rapid improvement of people's material living standard and the loss of spiritual pursuit or the increase of pressure is becoming more and more serious.With the impact of the new corona-virus pandemic, the employment situation of college students is more severe. This leads to the growth of psychological problems and alienation behavior of college students. In the promoting positive psychology and enhancing the happiness of the whole people, the relationship between interpersonal relationship and subjective well-being of college students is an interesting and valuable research topic.

Methods: The structural model with AMOS21.0 was used to verify the mediating effect of psychological capital between interpersonal relationship distress and SWB from the survey data of 673 college students.

Results: The interpersonal distress has a negatively correlated with psychological capital and SWB, while psychological capital has a positively correlated with SWB. Psychological capital has Partial mediating effect between interpersonal distress and SWB, and the indirect effect accounts for 57% of the total effect.

Conclusion: This study indicates that schools should pay attention to education and intervention in psychological capital, except for strengthening knowledge and skill training of college students in how to communicate with others.

Open-source communities(OSCs) are gaining significant attention in the current business environment of information technology(IT). More and more IT companies and individuals are exploring how to achieve innovation through open-source collaboration, and value co-creation(VCC) in the OSCs has become a trend. Therefore, it is particularly important to examine the mechanism of OSCs under the background of VCC theory. This study proposes a conceptual framework of open-source value co-creation (OSVCC), which is characterized by openness, sharing, collaboration, and freedom, for understanding the internal mechanisms and contextual conditions in the relationship between OSCs participants. This study constructed a pairwise combined four-category classification model combining the perspectives of the commercialization level (low and high) and the maturity stage (developmental and mature) of the OSCs. Based on the model, this study selects and analyzes four presentive cases of OSCs using a multiple case study approach. Then, this study proposes a framework for OSVCC to identify the crucial factors that promote the successful implementation of innovation and value creation. The OSVCC framework encompasses three primary participants, effective VCC processes, and key open-source principles. This study offers valuable managerial implications for enterprises that plan to participate in OSCs.

Samples from two outcrop sections, MGS1 and DGS1 of Milanggouwan and Dishaogouwan in the Salawusu River Basin, were studied in terms of grain size using end-member model. Results show that: 1) MGS1 layer particles are more concentrated, better sorting, and smaller skewness and kurtosis values than those of DGS1. Whereas in the upper part of the DGS1 section, the grain size of the paleodune is coarser, with better sorting and sharper peak, comparing with the lower lacustrine sediments. 2) Three end-member components, EM1 (end-member 1), EM2 (end-member 2) and EM3 (end-member 3), which reflect sedimentary dynamic characteristics, are extracted by end-member analysis. The EM1 indicates the hydrodynamic force with great variation, EM2 indicates transporting force by flowing water and EM3 indicates the depositional environment closely related to the wind activity. 3) According to the accumulation processes of MGS1 and DGS1 strata, a total of four climate periods can be identified, namely early warming period, Holocene peak period, fluctuating transition to cold period and unstable cooling period. Moreover, EM1 of MGS1 and DGS1 is basically consistent with both the sea surface temperature (SST) in the western tropical Pacific and global temperature trends during the Holocene, suggesting that the environmental fluctuations recorded by MGS1 and DGS1 can be correlated with each other.

A significant population of biocontrol microorganisms resides in the rhizosphere of plants, which can be utilized for plant disease control. To explore the potential of rhizosphere soil microorganisms as biocontrol agents against pepper blight, a bacterial strain Pa608 was screened from rhizosphere soil of pepper and identified as Pseudomonas aeruginosa through morphological characteristics and 16S rRNA sequences. The result showed that the strain Pa608 demonstrated antagonistic activity against Phytophthora capsici, effectively suppressing mycelial growth. The potted experiment showed a high control efficacy of 88.0%. Remarkably, the strain Pa608 also reduced the disease index of pepper blight in the field, resulting in control efficiencies of 74.9%. Moreover, the strain Pa608 also enhanced pepper plant height and yield. GC-MS analysis revealed the production of numerous secondary metabolites by the strain Pa608, with α-pinene displaying potent anti-oomycete activity by inhibiting P. capsici growth. In conclusion, P. aeruginosa Pa608 exhibited high biocontrol activity against P. capsici and can be utilized for the management of P. capsici in pepper cultivation.

The objective of this study is to determine the relationship between short-term temperature variability on neighboring days and mortality. The change in maximum temperature in Northern Virginia, Richmond, Roanoke, and Norfolk, Virginia, on neighboring days was calculated from airport observations and associated with total mortality over a multi-county area surrounding each weather station. The association between day-to-day temperature change and mortality, lagged over a 28-day period, was analyzed using distributed lag non-linear models that controlled for air quality, temporal trends, and other factors. Days following large temperature declines were associated with an increased risk of mortality in three of the four locations, and temperature increases were linked to higher mortality risk in two cities. For example, the relative risk of mortality for a 12°C daily temperature decline (1st percentile) was 1.74 [0.92, 3.27] in Roanoke and 1.16 [0.70, 1.92] in Richmond. The net effect of short-term temperature increases was smaller, with the largest relative risk of 1.03 [0.58, 1.83] for a 12°C increase (99th percentile) in maximum temperature in Norfolk. In Richmond and Roanoke, there was an observed lagged effect of increased mortality (maximum relative risks varying from 1.08 to 1.10) that extended from 5 to 25 days associated with large temperature declines of 15°C or more. In contrast, there was a strong and immediate (lag 0-3 day) increase in the risk of mortality (1.10 to 1.15) in northern Virginia and Norfolk when the temperature increase exceeded 10°C (short-term warming). In general, consecutive day warming had a more immediate mortality impact than short-term cooling, when the peak mortality is lagged by one week or more. However, cooling of at least 10°C after a hot (summer) day reduced mortality relative to comparable cooling following a cold (winter) day, which is associated with high mortality. This differential mortality response as a function of temperature suggests that there is some relationship between average temperature, temperature variability, and season. The findings of this study may be useful to public health officials in developing mitigation strategies to reduce the adverse health risks associated with short-term temperature variability.

In this study, we investigated the neuroprotective effect of a water extract of ginseng (WEG) obtained via low-temperature extraction of the brain of mice with Parkinson's disease (PD) and the ameliorative effect on the damaged intestinal system for the treatment of dyskinesia in PD mice. MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) was injected intraperitoneally into male C57BL/6 mice to establish a PD model, and WEG was given via oral gavage. The results indicated that WEG could protect the damaged neuronal cells of the mice brain, inhibit the aggregation of α-synuclein (α-Syn) in the brain, and increase the positive expression rate of tyrosine hydroxylase (TH). WEG significantly improved intestinal damage and regulated intestinal disorders (P<0.05). WEG intervention increased the levels of beneficial bacteria, such as Lactobacillus, and normalized the abundance and diversity of colonies in the intestine of mice. Our results suggested that WEG protected neurons in the brain of PD mice via inhibiting the aggregation of α-Syn in the brain and increasing the positive expression level of TH in the brain. WEG regulated the gut microbiota of mice, improved the behavioral disorders of PD mice, and offered some therapeutic effects on PD mice.

Modern medical technology innovation is a critical safeguard for human health, while a significant number of developing nations are confronted with the challenge of biopharmaceutical technological advancement. To investigate the possible routes of technological advancement, we investigated the impact of the endogenous knowledge spillover effect on firm innovation endeavors. Our research involves a theoretical two-stage R&D game model that is built around the characteristics of pharmaceutical organizations. Theoretical studies elucidated the impact as well as the probable mechanism of the spillover effect. To verify the theoretical study, we conducted econometric analysis using data from the pharmaceutical sector of Chinese enterprise listed on the A-share market. The study's findings indicate that endogenous knowledge spillovers impede organizations' innovation endeavors. This phenomenon may be attributed to the existence of the patent race paradigm and high concentration of enterprises' R&D endeavors in specific areas. Additional examination of heterogeneity demonstrates that private firms, small and medium-sized enterprises (SMEs), and non-high-tech enterprises experience a larger adverse impact from the spillover effect. Hence, we suggest implementing "loser's subsidies", reallocating R&D resources, and making modifications to competition policies as measures to enhance the innovation performance of biopharmaceutical markets. These policies will facilitate the technical advancement of medicines in developing nations.

In recent years, the surge in reviews and comments on newspapers and social media has made sentiment analysis a focal point of interest for researchers. Sentiment analysis is also gaining popularity in the Bengali language. However, Aspect-Based Sentiment Analysis is considered a difficult task in the Bengali language due to the shortage of perfectly labeled datasets and the complex variations in the Bengali language. This study used two open-source benchmark datasets of the Bengali language, Cricket, and Restaurant, for our Aspect-Based Sentiment Analysis task. The original work was based on the Random Forest, Support Vector Machine, K-Nearest Neighbors, and Convolutional Neural Network models. In this work, we used the Bidirectional Encoder Representations from Transformers, the Robustly Optimized BERT Approach, and our proposed hybrid transformative Random Forest and Bidirectional Encoder Representations from Transformers (tRF-BERT) models to compare the results with the existing work. After comparing the results, we can clearly see that all the models used in our work achieved better results than any of the previous works on the same dataset. Amongst them, our proposed transformative Random Forest and Bidirectional Encoder Representations from Transformers achieved the highest F1 score and accuracy. The accuracy and F1 score of aspect detection for the Cricket dataset were 0.89 and 0.85, respectively, and for the Restaurant dataset were 0.92 and 0.89 respectively.