A5100734399- Data Quality and Management
- Hydrology and Watershed Management Studies
- Privacy-Preserving Technologies in Data
- Data Mining Algorithms and Applications
- Hydrological Forecasting Using AI
- Groundwater flow and contamination studies
- Data Management and Algorithms
- Anomaly Detection Techniques and Applications
- Complexity and Algorithms in Graphs
- RNA Research and Splicing
- Flood Risk Assessment and Management
- Soil Moisture and Remote Sensing
- Web Data Mining and Analysis
- Interconnection Networks and Systems
- Topic Modeling
- RNA modifications and cancer
- Climate variability and models
- Domain Adaptation and Few-Shot Learning
- Advanced Graph Theory Research
- interferon and immune responses
- Water resources management and optimization
- Sentiment Analysis and Opinion Mining
- Grey System Theory Applications
- Game Theory and Voting Systems
- Recommender Systems and Techniques
Wuhan University
2025
Yunnan Normal University
2020-2025
Jilin Agricultural University
2022-2024
Nankai University
2020-2024
Tianjin International Joint Academy of Biomedicine
2023
Tsinghua University
2009-2022
Zhejiang University
2022
Baidu (China)
2022
Ministry of Education of the People's Republic of China
2022
Second Affiliated Hospital of Zhejiang University
2022
Abstract In this study, the authors applied version 4 of Community Land Model (CLM4) integrated with an uncertainty quantification (UQ) framework to 20 selected watersheds from Parameter Estimation Experiment (MOPEX) spanning a wide range climate and site conditions investigate sensitivity runoff simulations major hydrologic parameters assess fidelity CLM4, as land component Earth System (CESM), in capturing realistic hydrological responses. They found that for simulations, most significant...
Incomplete information often occurs along with many database applications, e.g., in data integration, cleaning, or exchange. The idea of imputation is to fill the missing values its neighbors who share same/similar information. Such could either be identified certainly by editing rules extensively similarity relationships. Owing sparsity, number w.r.t. value equality rather limited, especially presence variances. To enrich candidates, a natural consider relationship. However, candidates...
As a critical node for insulin/IGF signaling, insulin receptor substrate 1 (IRS-1) is essential metabolic regulation. A long and unstructured C-terminal region of IRS-1 recruits downstream effectors promoting signals. However, the underlying molecular basis this remains elusive. Here, we found that C-terminus undergoes liquid-liquid phase separation (LLPS). Both electrostatic hydrophobic interactions were seen to drive LLPS. Self-association IRS-1, which was mainly mediated by 301-600...
The main black land conservation measure in China is the straw return to fields. processing of high-resolution images collected by aerial photography UAVs through image stitching technology can provide information for achieving fast and accurate detection cover over large areas. classical SIFT algorithm has many drawbacks, such as high dimensionality feature descriptors, computational effort, low matching efficiency. To solve problems above, this study proposes an improved algorithm. First,...
Hepatocellular carcinoma (HCC) is one of the leading causes death, which deserves further study to reveal underlying molecular mechanisms. Studies have shown that miR-9 in associated with poor prognosis HCC patients. However, mechanisms transcriptional activation regulation and its role malignant progression been rarely investigated. Some coactivators can form phase-separated condensates at super-enhancers compartmentalize concentrate transcription apparatus drive robust gene expression....
Missing numerical values are prevalent, e.g., owing to unreliable sensor reading, collection and transmission among heterogeneous sources. Unlike categorized data imputation over a limited domain, the suffer from two issues: (1) sparsity problem, incomplete tuple may not have sufficient complete neighbors sharing same/similar for imputation, (almost) infinite domain; (2) heterogeneity different tuples fit same (regression) model. In this study, enlightened by conditional dependencies that...
Abstract. Realistically representing spatial heterogeneity and lateral land surface processes within between modeling units in Earth system models is important because of their implications to energy water exchanges. The traditional approach using regular grids as computational may lead inadequate representation subgrid movements water, carbon fluxes. Here a subbasin-based framework introduced the Community Land Model (CLM), which component System (CESM). Local are represented each subbasin...
Obesity is a leading risk factor for development of hepatocellular carcinoma (HCC). High-fat intake produces cytotoxic effects in liver cells, such as excessive reactive oxygen species (ROS) accumulation and apoptosis. How HCC cells regulate ROS level escape the high fat diet (HFD) stress remains unclear. Herein, this work reports critical anti-ROS/apoptotic role ubiquitin-like protein interferon stimulated gene 15 (ISG15) HFD-promoted HCC. In mouse models clinical samples, upregulation...
In order to meet the urgent need of fruit contour information for robot precision picking in complex field environments (such as light changes, occlusion and overlap, etc.), this paper proposes an improved YOLOv8s-seg method tomato instance segmentation, named ACP-Tomato-Seg. The two innovative modules: Adaptive Oriented Feature Refinement module (AOFRM) Custom Multi-scale Pooling (CMPRD) with Residuals Depth. By deformable convolution multi-directional asymmetric convolution, AOFRM...
Abstract. This study demonstrates the possibility of inverting hydrologic parameters using surface flux and runoff observations in version 4 Community Land Model (CLM4). Previous studies showed that calculations are sensitive to major CLM4 over different watersheds, illustrated necessity parameter calibration. Both deterministic least-square fitting stochastic Markov-chain Monte Carlo (MCMC)-Bayesian inversion approaches evaluated by applying them at selected sites with climate soil...
In the calibration of hydrological models, evaluation criteria are explicitly and quantitatively defined as single- or multi-objective functions when utilizing automatic approaches. most previous studies, there is a general opinion that no single-objective function can represent all important characteristics even one specific variable (e.g., streamflow). Thus hydrologists must turn to calibration. this study, we demonstrated an optimized compromise multi-response modes (i.e., functions)...
Relational graph neural networks have garnered particular attention to encode context in knowledge graphs (KGs). Although they achieved competitive performance on small KGs, how efficiently and effectively utilize for large KGs remains an open problem. To this end, we propose the Relation-based Embedding Propagation (REP) method. It is a post-processing technique adapt pre-trained KG embeddings with context. As relations are directional, model incoming head outgoing tail separately....
Misplaced data in a tuple are prevalent, e.g., value "Passport" is misplaced the passenger-name attribute, which should belong to travel-document attribute instead. While repairing in-attribute errors have been widely studied, i.e., repair error by other values domain, misplacement surprisingly untouched, where true simply some of same tuple. For instance, indeed record. In this sense, we need novel swapping model (to swap and "John Adam" tuple). Determining proper repair, however,...
Abstract In view of the nonlinear and linear influence vegetables in sales forecasting, previous single model could not fully explore variation law sales, a combined based on LightGBM LSTM is proposed, significant abstract characteristics affecting forecasting are explored respectively by combining advantages two models. First, intensive learning modeled analyzed, then models weighted array error reciprocal method for forecasting. The experimental results show that proposed combination more...
Missing values may appear in various attributes. By "various", we mean (1) different types of a tuple, such as numerical or categorical, and (2) attributes either the dependent determinant regression models dependency rules. Such varieties unfortunately prevent imputation performing. In this paper, propose to study distance that predict distances between tuples for missing data imputation. The immediate benefits are two aspects, uniformly processing collaboratively utilizing on all with...
The report from the 20th National Congress of Communist Party China (CPC) suggests advancing digitization education and establishing a society nation that prioritizes lifelong learning for all individuals. Higher serves as crucial platform nurturing innovative high-quality talent, playing vital role in modernizing education. With rise digital technology, music teaching methods colleges are encountering both new opportunities challenges. Currently, mode college programs, which fundamental...
Although clustering methods have shown promising performance in various applications, they cannot effectively handle incomplete data. Existing studies often impute missing values first before analysis and conduct these two processes separately. However, inaccurate imputation does not necessarily contribute positively to the subsequent clustering. Intuitively, accurate can serve benefit from each other, where clustering-based typically utilize cluster signals data fillings are expected bring...
Precise weed recognition is an important step towards achieving intelligent agriculture. In this paper, a novel model, Cotton Weed-YOLO, proposed to improve the accuracy and efficiency of detection. CW-YOLO based on YOLOv8 introduces dual-branch structure combining Vision Transformer Convolutional Neural Network address problems small receptive field CNN high computational complexity transformer. The Receptive Field Enhancement (RFE) module enable feature pyramid network adapt information...
With the growth of depth neural networks and scale data, difficulty network training also increases. When GPU memory is insufficient, it challenging to train deeper models. Recent research uses tensor swapping recomputation techniques in a combined manner optimize usage. However, complex dependencies enormous scales DNN graph limit improvement single optimization. Improper swap decisions even bring negative effects on performance. In this article, we propose novel hybrid re-generation...