A5100638959- Advanced Computational Techniques and Applications
- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Anomaly Detection Techniques and Applications
- Topic Modeling
- Service-Oriented Architecture and Web Services
- Complex Network Analysis Techniques
- Machine Learning in Bioinformatics
- Advanced Sensor and Energy Harvesting Materials
- Protein Structure and Dynamics
- Advanced Text Analysis Techniques
- MXene and MAX Phase Materials
- Artificial Immune Systems Applications
- Conducting polymers and applications
- E-commerce and Technology Innovations
- Remote Sensing and Land Use
- Bioinformatics and Genomic Networks
- Image and Object Detection Techniques
- Spectroscopy and Chemometric Analyses
- Network Security and Intrusion Detection
- Imbalanced Data Classification Techniques
- Diverse Aspects of Tourism Research
- Domain Adaptation and Few-Shot Learning
- Natural Language Processing Techniques
- Power Systems and Renewable Energy
Xinjiang University
2020-2025
Institute of Software
2025
State Key Laboratory of Software Engineering
2025
Zhejiang Ocean University
2025
Intelligent Software Solutions (United States)
2025
Fuzhou University
2008-2024
University of California, San Francisco
2024
South China Agricultural University
2024
Henan University
2024
Beijing Institute of Technology
2023
Protein function prediction is essential for disease treatment and drug development; yet, traditional biological experimental methods are less efficient in annotating protein function, existing automated fail to fully leverage multisource data. Here, we present MSF-PFP, a computational framework that fuses data features predict with high accuracy. Our designs specific models feature extraction based on the characteristics of various sources, including global-local-individual strategy local...
Abstract Due to the rotary workpiece on transmission mechanism in welding automatic production line entering next process from a certain process, rotation offset will occur along axis direction, resulting inaccurate positioning, which turn reduces surface machining accuracy of body. Hence, this paper proposes positioning method based visual recognition. Firstly, industrial monocular array camera is used collect image rotating body, and feature information identified extracted calibrate...
Proteins, as the fundamental macromolecules of life, play critical roles in various biological processes. Recent advancements intelligent protein function prediction methods leverage sequences, structures, and biomedical literature data. Among them, for sequences remain an enduring popular research direction. Existing studies have failed to effectively utilize multilevel attribute features reflected sequences. This limitation hinders enrichment descriptions needed high-precision functions....
Protein functional studies are crucial in the fields of drug target discovery and design. However, existing methods have significant bottlenecks utilizing multisource data fusion Gene Ontology (GO) hierarchy. To this end, study innovatively proposes DeepMFFGO model designed for protein function prediction under large-scale multifeature fusion. A fine-tuning strategy using intermediate-level feature selection is proposed to reduce redundancy sequences mitigate distortion top-level features....
To address the problem of inadequate feature extraction and binary code discrete optimization faced by deep hashing methods using a relaxation-quantization strategy, novel attention-guided method with pairwise labels (DAHP) is proposed to enhance global fusion, better learn contextual information image features effectively representation, solve losing in optimizing loss function. First, we introduce new concept called anchor hash generation(AHCG) algorithm, train ResNet position attention...
Abstract Solar‐powered interfacial water evaporation is a promising technique for alleviating freshwater stress. However, the performance of solar evaporators still constrained by low photothermal conversion efficiency and high enthalpy. Herein, 0D carbon quantum dots (CQDs) are combined with 2D MXene to serve as hybrid material enhance light absorption ability, meanwhile sodium carboxymethyl cellulose (CMC)/polyacrylamide (PAM) hydrogels used substrate transport reduce enthalpy evaporation....
Outlier detection is an important research direction in the field of data mining. Aiming at problem unstable results and low efficiency caused by randomly dividing features set Isolation Forest algorithm outlier detection, CIIF (Cluster-based Improved Forest) that combines clustering proposed. first uses k-means method to cluster set, selects a specific construct selection matrix based on clustering, implements mechanism through matrix; then builds multiple isolation trees. Finally, outliers...
Camera surveillance is widely used in residential areas, highways, schools and other public places. The monitoring scanning of sudden abnormal events depend on humans. Human anomaly not only consumes a lot manpower time but also has large error detection. Video detection based AE (Auto-Encoder) currently the dominant research approach. model highly symmetrical network structure encoding decoding stages. trained by learning standard video sequences, anomalous are later determined terms...
Purpose Incorporating dynamic spatial effects exhibits considerable potential in improving the accuracy of forecasting tourism demands. This study aims to propose an innovative deep learning model for capturing effects. Design/methodology/approach A novel founded on transformer architecture, called spatiotemporal network, is presented. has three components: temporal transformer, and fusion modules. The dependencies each attraction are extracted efficiently by module. correlations between...
Neural headline generation models have recently shown great results since neural network methods been applied to text summarization. In this paper, we focus on news generation. We propose a model based generative pre-training model. our model, rich features input module. The only contains decoder incorporating the pointer mechanism and n-gram language features, while other use encoder-decoder architecture. Experiments datasets show that achieves comparable in field of
In most of the existing multi-task learning (MTL) models, multiple tasks' public information is learned by sharing parameters across hidden layers, such as hard sharing, soft and hierarchical sharing. One promising approach to introduce model pruning into learning, sparse which regarded being outstanding in knowledge transferring. However, above method performs inefficiently conflict tasks, with inadequate private information, or through suffering from negative this paper, we propose a...
Depression is a prevalent mental disorder that poses significant risks to human health and social stability. Current methods for diagnosing depression heavily rely on patient descriptions psychiatrist observations, which are susceptible interference from subjective factors carry the risk of misdiagnosis missed diagnosis. Therefore, it crucial develop an objective method recognizing based criteria. Recently, combining EEG signals with deep learning techniques recognition has become popular...