A5052104124- Rough Sets and Fuzzy Logic
- Topic Modeling
- Data Mining Algorithms and Applications
- Advanced Image and Video Retrieval Techniques
- Text and Document Classification Technologies
- Natural Language Processing Techniques
- Advanced Computational Techniques and Applications
- Video Surveillance and Tracking Methods
- Advanced Text Analysis Techniques
- Biomedical Text Mining and Ontologies
- Machine Learning in Healthcare
- Computational Drug Discovery Methods
- Image Retrieval and Classification Techniques
- Web Data Mining and Analysis
- Advanced Neural Network Applications
- Data Management and Algorithms
- Multimodal Machine Learning Applications
- Anomaly Detection Techniques and Applications
- Recommender Systems and Techniques
- Image and Signal Denoising Methods
- Human Pose and Action Recognition
- Advanced Steganography and Watermarking Techniques
- Face and Expression Recognition
- Multi-Criteria Decision Making
- Advanced Image Processing Techniques
Dalian Maritime University
2016-2025
Heilongjiang University
2024
Beijing Institute of Radio Metrology and Measurement
2024
Dalian University of Technology
2023
Beijing Jiaotong University
2013-2023
Shandong Jianzhu University
2023
Anhui Sanlian University
2021-2022
Massachusetts Institute of Technology
2022
Xi’an Jiaotong-Liverpool University
2021
Tsinghua University
2002-2020
Medicine recommendation aims to provide a combination of medicine based on the patient's electronic health record (EHR), which is an essential task in healthcare. Existing methods either base recommendations EHRs or models with knowledge drug–drug interactions (DDIs) achieve DDI reduction. However, former history but ignores undesirable DDIs, while latter lacks mining patient records and gets low accuracy. Therefore, this study contributes research personalized medication that consider drug...
Biomedical named entity recognition (BNER) is the basis of biomedical text mining and one core sub-tasks information extraction. Previous BNER models based on conventional machine learning rely time-consuming feature engineering. Though most neural network methods improve problems with automatic learning, they cannot pay attention to significant areas when capturing features. In this paper, we propose an attention-based BiLSTM-CRF model. First, model adopts a bidirectional long short-term...
Human parsing has attracted considerable research interest due to its broad potential applications in the computer vision community. In this paper, we explore several useful properties, including high-resolution representation, auxiliary guidance, and model robustness, which collectively contribute a novel method for accurate human both simple complex scenes. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Starting from scenes</i> : propose...
Image feature matching, a foundational task in computer vision, remains challenging for multimodal image applications, often necessitating intricate training on specific datasets. In this paper, we introduce Unified Feature Matching pre-trained model (UFM) designed to address matching challenges across wide spectrum of modal images. We present Multimodal Assistant (MIA) transformers, finely tunable structures adept at handling diverse problems. UFM exhibits versatility addressing both tasks...
Fundus images are one of the main methods for diagnosing eye diseases in modern medicine.The vascular segmentation fundus is an essential step quantitative disease analysis.Based on previous studies, we found that category imbalance reasons restrict improvement accuracy.This paper presents a new method supervised retinal vessel can effectively solve above problems.In recent years, it popular using deep learning to segmentation.We have improved loss function order better handle imbalances.By...
Image matching between the optical and synthetic aperture radar (SAR) is one of most fundamental problems for earth observation. In recent years, many researchers have used hand-made descriptors with their expertise to find matches SAR images. However, due large nonlinear radiation difference images images, image becomes very difficult. To deal problems, article proposes an efficient feature position algorithm (MatchosNet) based on local deep descriptor. First, A new dataset presented by...
Traditional drug development is often high-risk and time-consuming. A promising alternative to reuse or relocate approved drugs. Recently, some methods based on graph representation learning have started be used for repositioning. These models learn the low dimensional embeddings of disease nodes from drug-disease interaction network predict potential association between drugs diseases. However, these strict requirements dataset, if dataset sparse, performance will severely affected. At same...
Pietrain swine homozygous for the hal gene (n) associated with porcine stress syndrome (PSS) and a Pietrain-derivative breed, Near (NP), frequency of .31 n, were mated to produce reciprocal F1, F2, purebred NP litters. The halothane challenge test was performed on all 40 parents 240 progeny predict their genotype PSS susceptibility. DNA-based assay C T mutation at base pair 1,843 skeletal muscle ryanodine receptor (ryr1) cDNA, which is very highly correlated PSS, also determined these...
To solve the problem of semantic loss in text representation, this paper proposes a new embedding method word representation space called wt2svec based on supervised latent Dirichlet allocation (SLDA) and Word2vec. It generates global topic vector utilizing SLDA which can discover information through topics whole document set. gets local The is obtained by combining with information. Additionally, named doc2svec generated. experimental results different datasets show that model obviously...