Xiaoli Wang

The increasing frequency of extreme weather events has resulted in a rapid growth of unstructured meteorological texts, posing challenges for accurate disaster-related information extraction. Existing relation extraction models that integrate semantic and syntactic information via Graph Convolutional Networks (GCNs) often suffer from noise amplification and information redundancy. To address these limitations, we propose a domain-specific relation extraction framework that integrates Bidirectional Encoder Representations from Transformers (BERT) with an Adaptive Denoising Graph Network. A dual-tuning architecture is designed to jointly model semantic and syntactic features. Specifically, a Self-Attention Graph Convolutional Network (SA-GCN) dynamically prunes irrelevant dependency structures, while a Residual Shrinkage Network (RS-Net) performs fine-grained feature denoising through sample-adaptive threshold learning. Experiments on the public DUIE 2.0 dataset and a newly constructed Agricultural Meteorological Disaster dataset demonstrate that the proposed model achieves F1 scores of 91.51% and 92.84%, respectively, outperforming state-of-the-art methods. Ablation studies further confirm the effectiveness of adaptive denoising in handling complex linguistic structures in disaster reports. The proposed approach enhances the robustness and accuracy of meteorological relation extraction and provides technical support for knowledge graph construction and intelligent disaster early warning systems.

Genomic prediction (GP) plays a pivotal role in expediting genetic gains for crop breeding. Recently, deep learning models has attracted growing attention in this field due to its superior performance over traditional models. However, most advanced models rely on simplified numerical representations of genomic variants and treat individuals in isolation, failing to capture the full complexity of genetic variation and the genetic relatedness among individuals. In response, we propose RAGP, a deep learning model driven by a retrieval-augmented mechanism composed of two synergistic components: 1) an embedding-based retrieval module that identifies a group of nearest neighbors highly relevant to the target sample as references, employs a gene-specific discriminator to refine the representation space and capture informative genetic structures; 2) an augmentation module that integrates the retrieved references to enhance the target representation. The enriched representation is then propagated through a regression network for final prediction. This retrieval-guided strategy enables more informed and fine-grained utilization of genomic data and strengthens the model’s capacity to capture individual-level genetic relationships. Extensive experiments reveal that RAGP achieves consistently superior predictive performance compared to both conventional methods and state-of-the-art deep learning models. By incorporating retrieval-augmented mechanisms, RAGP achieves substantial improvements in predictive accuracy and offers a promising direction for advancing genomic selection in crop breeding. Our code is available on https://github.com/l00907l/RAGP.

Based on a number of field surveys and studies on the key wetlands in Southern Xinjiang for one year, we obtained the original data of plant species information. And then we supplemented, improved and proofread the original data using relevant information from various books about biodiversity and flora as well as related professional databases such as the Flora of China. Following the APGIV classification system (2016), we sorted out, reviewed, and examined the original data, so as to obtain a dataset of plant lists from key surveyed. This dataset covers 35 key survey wetlands, including plant species name, Latin name, alias, phylum, class, order, Chinese family name, Latin family name, Chinese genus name, Latin genus name, distributed wetland name, habitat, plant value and other information. The dataset comprises 11 orders, 18 families, 33 genera, and 58 species. This dataset can provide data support for plant species research, ecological environment quality restoration and rational utilization, as well as the biodiversity protection research in the wetland region of southern Xinjiang in the future.

In this study, we collected fifteen species of Salsola L. from Northern Xinjiang at the end of the flowering period in 2007, including Salsola. subcrassa, S. orientalis, and S. lanata, etc.. We first identified and classified these species by comparison of flowers and fruits, and then photographed the pollen from multiple angles, and analyzed them a using a LEO-1430VP scanning electron microscope with a maximum magnification of 20,000× and an image resolution of 1023×766 pixels. A total of 84 images were included in the study. The Salsola L. in the dataset was identified twice by an expert in Chenopodiaceae taxonomy, and a specialized staff operated a scanning electron microscope to observe pollen traits. Readers can access the corresponding pollen scanning electron microscope images in this dataset by referecing the species name of Salsola L.. The release of this dataset facilitates the in-depth study of the typical pollen characteristics of Salsola L. in Xinjiang and even in Central Asia. Moreover, it further complements and perfects the grouping system based on the morphological characteristics of Salsola L., and provides an important reference for the research on palynology, taxonomy, ecology and so on.

Solanaceous vegetables are popular dishes for residents, and they are also one type of the important cultivated vegetables in China. The price of solanaceous vegetables is non-steady with large fluctuations. Therefore, it is important to master the historical price trends and fluctuation laws, which can provide scientific guidance and support for ensuring the stability of vegetable supply prices and guiding the orderly operation of the market. Through data collection, stability test and HP filtering method decomposition, we formed a dataset of the HP filter model-based research on price volatility of solanaceous vegetables at Beijing Xinfadi Agricultural and Sideline Products Wholesale Market from 2012 to 2018. The varieties cover five kinds of solanaceous vegetables (eggplant, tomato, green pepper, bean and cucumber), including market, date, weekly/monthly average price, weekly/monthly average volatility, weekly/monthly maximum price, weekly/monthly minimum price, weekly/monthly average HP filter trend value, weekly/monthly average HP filter fluctuation value. The dataset can reflect the weekly and monthly changes of the trading prices of solanaceous vegetables in Beijing Xinfadi Market, and provide scientific data foundation for the research on as well as monitoring and early warning of the price fluctuation of solanaceous vegetables in Beijing.

Xinjiang is rich in wild medicinal resources, which are mainly distributed in Tianshan Mountains and Altai Mountains in Northern Xinjiang. Trollius chinensis is one of the plants with heat clearing and detoxifying effects. It has been listed as one of the recommended Chinese medicinal materials for poverty alleviation in China’s traditional Chinese medicine industry, with a good market prospect. This dataset is formed based on the field study on Trollius chinensis in Altay, Xinjiang. It consists of three types of multi-source data: 81 Trollius chinensis sample point data, 42 environmental variable data and one Trollius chinensis distribution data. Among them, the data of sample points were collected through field study in July, 2018 as well as June and July, 2019; The number of environmental variables includes 39 climatic data and 3 topographic data. The climatic data are obtained by interpolating the data of 6 meteorological stations in Altay region from 1995 to 2018, and the topographic data are obtained based on elevation data processing; the distribution data of Trollius chinensis are obtained through inputting the sample point data and environmental variable data into MaxEnt model for analysis. This dataset can provide a data basis for predicting the function evaluation of the potential suitable growth area of Altay Golden Lotus. Referring to the distribution area of Altay Golden Lotus, suitable growth areas can be selected for introduction and artificial cultivation. The dataset can support the resource development and species diversity protection of Altay wild Golden chinensis, and prop up the research of scientific researchers on such medicinal plants. Moreover, it also has certain value for the scientific research of other wild medicinal plants in this field.