Search

Accurate and cost-effective quantification of the carbon cycle for agroecosystems at decision-relevant scales is critical to mitigating climate change and ensuring sustainable food production. However, conventional process-based or data-driven modeling approaches alone have large prediction uncertainties due to the complex biogeochemical processes to model and the lack of observations to constrain many key state and flux variables. Here we propose a Knowledge-Guided Machine Learning (KGML) framework that addresses the above challenges by integrating knowledge embedded in a process-based model, high-resolution remote sensing observations, and machine learning (ML) techniques. Using the U.S. Corn Belt as a testbed, we demonstrate that KGML can outperform conventional process-based and black-box ML models in quantifying carbon cycle dynamics. Our high-resolution approach quantitatively reveals 86% more spatial detail of soil organic carbon changes than conventional coarse-resolution approaches. Moreover, we outline a protocol for improving KGML via various paths, which can be generalized to develop hybrid models to better predict complex earth system dynamics.

Relict subtropical coniferous forests in China face severe fragmentation, resulting in declining populations, and some are under significant threat from invasive alien species. Despite the crucial importance of understanding forest dynamics, knowledge gaps persist, particularly regarding the impact of invasive plants on vulnerable natives like Keteleeria evelyniana. In this study, we investigated the impact of invasive plants on the regeneration of forests dominated by K. evelyniana, a subtropical relict species in southwestern China. For this purpose, we characterized forest dynamics of 160 forest plots featuring K. evelyniana as the primary dominant species and determined whether the presence of invasive plants was correlated with regeneration of K. evelyniana. We identified four distinct forest types in which K. evelyniana was dominant. We found that radial growth of K. evelyniana trees is faster in younger age-classes today than it was for older trees at the same age. The population structure of K. evelyniana in each forest type exhibited a multimodal age-class distribution. However, three forest types lacked established saplings younger than 10 years old, a situation attributed to the dense coverage of the invasive alien Ageratina adenophora. This invasive species resulted in a reduction of understory species diversity. Additionally, our analysis uncovered a significant negative correlation in phylogenetic relatedness (net relatedness index) between native and invasive alien plant species in eastern Yunnan. This suggests closely related invasive species face heightened competition, hindering successful invasion. Taken together, our findings indicate that successful establishment and habitat restoration of K. evelyniana seedling/saplings require effective measures to control invasive plants.

Resident training is a necessary path to cultivate excellent clinical doctors. Based on the Consensus on Core Competency Framework for Residency Education Among China Consortium of Elite Teaching Hospitals for Residency Education, Department of Neurology from Peking Union Medical College Hospital has established a training program for neurology residents with a graded evaluation system and a progressive clinical responsibility concept guided by core competency. Overcoming the teaching difficulties of neurology and considering the clinical characteristics of neurology, we have established a series of replicable and promotable neurology residency training curriculum through inheritance and innovation, which provides reference for the continuous improvement of China's neurology residency training program.

Transformer is a deep learning model based on the self-attention mechanism, showing tremendous potential in computer vision. In image classification tasks, the key challenge lies in efficiently and accurately capturing both local and global features of input images. Traditional approaches rely on convolutional neural networks to extract local features at the lower layers, expanding the receptive field through stacked convolutional layers to obtain global features. However, this strategy aggregates information over relatively short distances, making it difficult to model long-term dependencies. In contrast, the self-attention mechanism of Transformer directly compares features across all spatial positions, capturing long-range dependencies at both local and global levels and exhibiting stronger global modeling capabilities. Therefore, a thorough exploration of the challenges faced by Transformer in image classification tasks is crucial. Taking Vision Transformer as an example, this paper provides a detailed overview of the core principles and architecture of Transformer. It then focuses on image classification tasks, summarizing key issues and recent advancements in visual Transformer research related to performance enhancement, computational costs, and training optimization. Furthermore, applications of Transformer in specific domains such as medical imagery, remote sensing, and agricultural images are summarized, highlighting its versatility and generality. Finally, a comprehensive analysis of the research progress in visual Transformer for image classification is presented, offering insights into future directions for the development of visual Transformer.

Objective： The effect of UV sterilization on the texture quality and flavor substances of the frozen grass carp meat was studied, and the best UV irradiation time was determined. Methods: The grass carp meat was pretreated with different UV irradiation times （0, 30, 60, 90, 150 min） at 30 W of UV radiation, then the quality parameters （apparent morphology, total number of bacteria on the surface of fish, juice loss rate, texture parameters, color, odor and sensory） and volatile components of the fish meat during chilling storage （0, 3, 6, 9 d） were measured. Results: It showed that the longer the UV irradiation time （0~150 min）, the better the bacteria reduction effect on the surface of fish, the lower the juice loss rate, and the later the putrefaction smell appeared. However, it had little effect on the texture, color of the fish, and the total bacteria during the latter of chilling storage （6~9 d）. In the detection of flavor substances, 90 volatile components were detected, mainly aldehydes, ketones, alcohols, esters, and acids. Octanal, nonanal, hexanal, 2,5-octadione, 1-hexanol, and 1-octen-3-alcohol were the main volatile components after UV treatment. In the later stage of cold storage, esters and acids were increased gradually, which might be related to the formation of the characteristic aroma of fish meat. Compared with the control group, UV irradiation time of 90 min showed the best comprehensive effect, and the quality was decreased with the extension of irradiation time. Conclusion: UV sterilization pretreatment can delay the deterioration of quality and the decline of flavor of refrigerated grass carp meat in the early stage of chilling storage but has little effect on the quality of grass carp meat in the later stage of chilling storage.