A5009773620- Single-cell and spatial transcriptomics
- Cell Image Analysis Techniques
- Gene Regulatory Network Analysis
- Human Pose and Action Recognition
- Extracellular vesicles in disease
- Osteoarthritis Treatment and Mechanisms
- Knee injuries and reconstruction techniques
- Anomaly Detection Techniques and Applications
- Context-Aware Activity Recognition Systems
- Traffic Prediction and Management Techniques
- Gene expression and cancer classification
- Artificial Intelligence in Law
- Law, Economics, and Judicial Systems
- Metallurgy and Material Forming
- Cardiac Fibrosis and Remodeling
- BIM and Construction Integration
- Immune cells in cancer
- Machine Learning and ELM
- Metal Forming Simulation Techniques
- Traffic control and management
- Congenital heart defects research
- Internet Traffic Analysis and Secure E-voting
- Advanced Surface Polishing Techniques
- European and International Contract Law
- Energy Load and Power Forecasting
Institute of Economics
2025
Tsinghua University
2018-2024
Carnegie Mellon University
2018
Institute of Bioinformatics
2018
Imperial College London
2016
Deep neural networks, including recurrent have been successfully applied to human activity recognition. Unfortunately, the final representation learned by networks might encode some noise (irrelevant signal components, unimportant sensor modalities, etc.). Besides, it is difficult interpret gain insight into models' behavior. To address these issues, we propose two attention models for recognition: temporal and attention. These mechanisms adaptively focus on important signals modalities....
The accumulation of massive single-cell omics data provides growing resources for building biomolecular atlases all cells human organs or the whole body. true assembly a cell atlas should be cell-centric rather than file-centric. We developed unified informatics framework seamless and built Ensemble Cell Atlas (hECA) from scattered data. hECA v1.0 assembled 1,093,299 labeled 116 published datasets, covering 38 11 systems. invented three new methods applications based on assembly: "in data"...
Idiopathic pulmonary fibrosis (IPF) is a chronic interstitial lung disease with high mortality rate and unclarified aetiology. Immune response elaborately regulated during the progression of IPF, but immune cells subsets are complicated which has not been detailed described IPF progression. Therefore, in current study, we sought to investigate role regulation by characterize heterogeneous IPF. To this end, performed single-cell profiling isolated from four stages bleomycin-induced fibrosis-a...
Abstract Profiling spatial variations of cellular composition and transcriptomic characteristics is important for understanding the physiology pathology tissues. Spatial transcriptomics (ST) data depict gene expression but currently dominating high-throughput technology yet not at single-cell resolution. Single-cell RNA-sequencing (SC) provide information level lack information. Integrating these two types would be ideal revealing landscapes We develop method STEM (SpaTially aware EMbedding)...
Large language models (LLMs) have demonstrated significant potential in transforming real estate transactions through advanced natural processing capabilities. This paper proposes a method leveraging LLMs to extract information from contracts and facilitate interactive querying by users. By emphasizing the use of few-shot learning, can effectively analyze contract documents respond specific questions users regarding details. explores development this method, its implementation workflows,...
Real estate sales contracts contain crucial information for property transactions, but manual extraction of data can be time-consuming and error-prone. This paper explores the application large language models, specifically transformer-based architectures, automated from real contracts. We discuss challenges, techniques, future directions in leveraging these models to improve efficiency accuracy contract analysis.
This article explores the application of Large Lan-guage Models (LLMs) and multi-task learning inevaluating validity contract formation sce-narios. The breaks down formationinto four essential elements: offer, acceptance, con-sideration, defenses then assesses pres-ence each element within a hypothetical case todetermine viability formation.
Background: Musculoskeletal tissue degeneration impairs the life quality and function of many people. Meniscus is a major origin knee osteoarthritis common threat to athletic ability, but its cellular mechanism remains elusive. Methods: We built cell atlas 12 healthy or degenerated human meniscus samples from inner outer meniscal zones 8 patients using scRNA-seq investigate microenvironment homeostasis changes in process verified findings with immunofluorescent imaging. Results: identified...
This perspective discusses the need and directions for development of a unified information framework to enable assembly cell atlases revolution in medical research on virtual body assembled systems.
A universal coordinate system that can ensemble the huge number of cells and capture their heterogeneities is vital importance for constructing large-scale cell atlases as references molecular cellular studies. Studies have shown exhibit multifaceted in transcriptomic features at multiple resolutions. This nature complexity makes it hard to design a fixed through combination known features. It desirable build learnable model major serve controlled generative data augmentation. We developed...
Understanding perturbations at the single-cell level is essential for unraveling cellular mechanisms and their implications in health disease. The growing availability of biological data has driven development a variety silico perturbation methods designed analysis, which offer means to address many inherent limitations experimental approaches. However, these computational are often tailored specific scenarios validated on limited datasets metrics, making evaluation comparison challenging....
The use of Finite Element (FE) simulation software to adequately predict the outcome sheet metal forming processes is crucial enhancing efficiency and lowering development time such processes, whilst reducing costs involved in trial-and-error prototyping. Recent focus on substitution steel components with aluminum alloy alternatives automotive aerospace sectors has increased need simulate behavior alloys for ever more complex component geometries. However these alloys, particular their high...
Expectations of machine learning (ML) are high for discovering new patterns in high-throughput biological data, but most such practices accustomed to relying on existing knowledge conditions design experiments. Investigations the power and limitation ML revealing complex from data without guide have been lacking. In this study, we conducted systematic experiments ab initio discovery with methods single-cell RNA-sequencing early embryonic development. Results showed that a strategy combining...
SUMMARY Single-cell omics data can characterize multifaceted features of massive cells and bring significant insights to biomedical researches. The accumulation single-cell provides growing resources for constructing atlases all a human organ or the whole body. true assembly cell atlas should be cell-centric rather than file-centric. We proposed unified information framework enabling seamless developed Ensemble Cell Atlas (hECA) as an instance. hECA version 1.0 assembled scRNA-seq across...
Abstract Profiling spatial variations of cellular composition and transcriptomic characteristics is important for understanding the physiology pathology tissues in health or diseases. Spatial transcriptomics (ST) data are powerful depicting gene expression but currently dominating high-throughput technology yet not at single-cell resolution. On other hand, RNA-sequencing (SC) provide information level lack information. Integrating these two types would be ideal revealing landscapes We...
Abstract A universal coordinate system that can ensemble the huge number of cells and capture their heterogeneities is vital importance for constructing large-scale cell atlases as references future molecular cellular studies. Studies have shown in complex organs exhibit multifaceted transcriptomic features at multiple resolutions. This nature complexity makes it hard to design a fixed through combination known features. It desirable build learnable model major serve controlled generative...
Abstract Multiple steps of bioinformatics processing are needed to convert the raw scRNA-seq data information that can be used in downstream analyses and building cell atlases. Dozens software packages have been developed different labs tend preferences on choices workflow. Such diversity cause difficulties future efforts aggregating from multiple labs, also for new start this field. A few pipelines help integrating into a whole, but fixed architecture makes it hard developers add features...
Deep neural networks, including recurrent have been successfully applied to human activity recognition. Unfortunately, the final representation learned by networks might encode some noise (irrelevant signal components, unimportant sensor modalities, etc.). Besides, it is difficult interpret gain insight into models' behavior. To address these issues, we propose two attention models for recognition: temporal and attention. These mechanisms adaptively focus on important signals modalities....
Abstract Background Musculoskeletal tissue degeneration impairs the life quality and function of many people. Meniscus is a major origin knee osteoarthritis common threat to athletic ability, but its cellular mechanism remains elusive. Methods We built cell atlas healthy/degenerated human meniscus using scRNA-seq investigate meniscal microenvironment homeostasis changes in process verified findings with immunofluorescent imaging. Results identified localized types inner outer meniscus, found...