A5102008661- Geotechnical Engineering and Underground Structures
- Railway Engineering and Dynamics
- RNA and protein synthesis mechanisms
- RNA modifications and cancer
- Machine Learning in Bioinformatics
- Infrastructure Maintenance and Monitoring
- Neonatal Respiratory Health Research
- Congenital Diaphragmatic Hernia Studies
- Structural Engineering and Vibration Analysis
- Mechanical stress and fatigue analysis
- Glycosylation and Glycoproteins Research
- Supramolecular Self-Assembly in Materials
- Civil and Geotechnical Engineering Research
- Advanced Proteomics Techniques and Applications
- RNA Research and Splicing
- Landslides and related hazards
- Digital Imaging for Blood Diseases
- Wireless Signal Modulation Classification
- Polydiacetylene-based materials and applications
- Geotechnical Engineering and Soil Mechanics
- Computational Drug Discovery Methods
- Bioinformatics and Genomic Networks
- Cell Image Analysis Techniques
- Geotechnical Engineering and Analysis
- Advanced biosensing and bioanalysis techniques
Fudan University
2021-2024
Tongji University
2020-2024
Beijing Academy of Artificial Intelligence
2024
Shanghai Artificial Intelligence Laboratory
2024
Artificial Intelligence Research Institute
2024
Guangzhou Metro Group (China)
2023
Guangzhou Metro Design & Research Institute
2023
North China Electric Power University
2023
Harbin Medical University
2022-2023
Northeastern University
2014
De novo peptide sequencing from mass spectrometry (MS) data is a critical task in proteomics research. Traditional de algorithms have encountered bottleneck accuracy due to the inherent complexity of data. While deep learning-based methods shown progress, they reduce problem translation task, potentially overlooking nuances between spectra and peptides. In our research, we present ContraNovo, pioneering algorithm that leverages contrastive learning extract relationship peptides incorporates...
Bronchopulmonary dysplasia (BPD) is a common chronic lung disease in extremely preterm neonates. The outcome and clinical burden vary dramatically according to severity. Although some prediction tools for BPD exist, they seldom pay attention severity are based on populations developed countries. This study aimed develop machine learning models selected factors Chinese population.
Predicting RNA secondary structures is crucial for understanding function, designing RNA-based therapeutics, and studying molecular interactions within cells. Existing deep-learning-based methods structure prediction have mainly focused on local structural properties, often overlooking the global characteristics evolutionary features of sequences. Guided by biological priors, we propose PriFold, incorporating two key innovations: 1) improving attention mechanism with pairing probabilities to...
The mitochondrial unfolded protein response (UPRmt) is a mitochondria stress response, which exerts crucial role in maintaining proteostasis during stress. However, there no bibliometric analyses systematically studied this field could comprehensively review research trends, evaluate publication performances and provide future perspectives.Articles investigating UPRmt published between 1994 2021 were downloaded from the Core Collection of Web Science (WOS). CiteSpace VOSviewer software...
Abstract RNA plays a pivotal role in translating genetic instructions into functional outcomes, underscoring its importance biological processes and disease mechanisms. Despite the emergence of numerous deep learning approaches for RNA, particularly universal language models, there remains significant lack standardized benchmarks to assess effectiveness these methods. In this study, we introduce first comprehensive benchmark BEACON ( BE nchm A rk CO mprehensive R N Task Language Models)....
Heavy rainfall has posed a great challenge to the service performance of high-speed rail (HSR) substructure, resulting in reduction ride quality and safety trains. To carry out proper repair work for it is imperative realize efficient identification precipitation-induced subgrade defects. this end, paper aims extract features typical defects from multiple track inspection data provide basis defect identification. Firstly, geotechnical site investigation including Ground Penetrating Radar...
Considering the discrete and stochastic properties of granular materials, a two-dimensional numerical model was established to study dynamic behavior high-speed railway (HSR) ballastless track subgrade by using finite difference-discrete element (FDM-DEM) coupling method. Combined with on-site vibration test results in macroscopic scale, mesoscopic response materials analyzed under different speed levels. Some unusual phenomena were observed as follow: 1). In surface layer (i.e., top area...
De novo peptide sequencing from mass spectrometry (MS) data is a critical task in proteomics research. Traditional de algorithms have encountered bottleneck accuracy due to the inherent complexity of data. While deep learning-based methods shown progress, they reduce problem translation task, potentially overlooking nuances between spectra and peptides. In our research, we present ContraNovo, pioneering algorithm that leverages contrastive learning extract relationship peptides incorporates...
In this paper, we focus on the Bronchopulmonary Dysplasia (BPD) prediction task, which aims to identify BPD in premature infants based given medical images. The existing methods for task sometimes learn spurious relations (confounders) between image and label while ignore causal features due limited size of dataset, largely influences interpretability robustness model. To address challenge problem, propose an interpretable method prediction, can eliminate irrelevant capture features. We term...
Abstract Automatic modulation classification (AMC) plays a vital role in modern communication systems, which can support wireless systems with limited spectrum resource. This paper proposes an AMC method, integrates gated recurrent unit (GRU) and convolutional neural network (CNN) to utilize the complementary input features of received signals for spatiotemporal feature extraction classification. Different from other state-of-the-art (SoA) frameworks, proposed classifier, named as fusion GRU...
Embedded turnout is one of the key facilities in embedded track network modern tram systems. In this study, damping performance and stiffness an are tested on site. A multi-rigid body dynamics model a finite-element established, two models coupled based wheel-rail contact theory for area. Next, dynamic responses passing through analyzed. The results show that has favorable performance. high variable section rails at switch frog would lead to increase force. When passes turnout, vertical...
Abstract Peptide sequencing via tandem mass spectrometry (MS/MS) is fundamental in proteomics data analysis, playing a pivotal role unraveling the complex world of proteins within biological systems. In contrast to conventional database searching methods, deep learning models excel de novo peptides absent from existing databases, thereby facilitating identification and analysis novel peptide sequences. Current for predominantly use an autoregressive generation approach, where early errors...
Understanding the three-dimensional structure of RNA is crucial for studying various biological processes. Accurate alignment and comparison structures are essential illustrating functionality evolution. The existing tools suffer from limitations such as size-dependency scoring functions inadequate handling short fragments, leading to conflicting interpretation structural functional relationships among molecules. Hence, we introduce RTM-align, a novel tool enhanced RNAs. RTM-align employs...
RNA plays a pivotal role in translating genetic instructions into functional outcomes, underscoring its importance biological processes and disease mechanisms. Despite the emergence of numerous deep learning approaches for RNA, particularly universal language models, there remains significant lack standardized benchmarks to assess effectiveness these methods. In this study, we introduce first comprehensive benchmark BEACON (\textbf{BE}nchm\textbf{A}rk \textbf{CO}mprehensive R\textbf{N}A Task...
As key elements within the central dogma, DNA, RNA, and proteins play crucial roles in maintaining life by guaranteeing accurate genetic expression implementation. Although research on these molecules has profoundly impacted fields like medicine, agriculture, industry, diversity of machine learning approaches-from traditional statistical methods to deep models large language models-poses challenges for researchers choosing most suitable specific tasks, especially cross-omics multi-omics...
Cell detection on immunohistochemistry stained (IHC-stained) images plays an essential role in computer assisted prediction of tumor progression and treatment response. Currently available cell datasets provide either point level or bounding box annotations for deep object network training. And these widely used networks usually employ standard pyramid structured multi-scale feature fusion. However, we find that methods have obvious limitation when facing large amounts cells similar scale...
Noise cancellation is the basic step of road surface image processing, such as roughness detection, crack detecting. The 2-D data will always be corrupted by noise in process sampling and digitalizing. Conventional filtering techniques median filter mean are designed to dispel single type noise. While contaminated mixed practical situation, because vibration disturbance. To solve problem we develop a mining approach for identification, further proposes fuzzy combined with characteristic...