A5100746427- Magnetic properties of thin films
- Magnetic Properties and Applications
- Higher Education and Teaching Methods
- Theoretical and Computational Physics
- Multimodal Machine Learning Applications
- Geographic Information Systems Studies
- Adversarial Robustness in Machine Learning
- Genetic Associations and Epidemiology
- Data Management and Algorithms
- Complexity and Algorithms in Graphs
- Anomaly Detection Techniques and Applications
- RNA modifications and cancer
- Advanced Database Systems and Queries
- Topic Modeling
- Time Series Analysis and Forecasting
- Advanced Text Analysis Techniques
- Human Pose and Action Recognition
- Radiomics and Machine Learning in Medical Imaging
- Ferroptosis and cancer prognosis
- Privacy-Preserving Technologies in Data
- Advanced Image and Video Retrieval Techniques
- Complex Network Analysis Techniques
- Explainable Artificial Intelligence (XAI)
- Human Mobility and Location-Based Analysis
- Low-power high-performance VLSI design
Nanjing Medical University
2024-2025
Florida International University
2014-2024
University of Jinan
2024
Washington State University
2024
University of Zurich
2023
Shanghai Institute of Technology
2023
Minzu University of China
2023
Amazon (United States)
2021-2022
Ministry of Industry and Information Technology
2022
Northwestern Polytechnical University
2022
Deep neural networks (DNNs) are an indispensable machine learning tool despite the difficulty of diagnosing what aspects a model’s input drive its decisions. In countless real-world domains, from legislation and law enforcement to healthcare, such diagnosis is essential ensure that DNN decisions driven by appropriate in context use. The development methods studies enabling explanation DNN’s has thus blossomed into active broad area research. field’s complexity exacerbated competing...
Deep neural networks (DNNs) have become a proven and indispensable machine learning tool. As black-box model, it remains difficult to diagnose what aspects of the model's input drive decisions DNN. In countless real-world domains, from legislation law enforcement healthcare, such diagnosis is essential ensure that DNN are driven by appropriate in context its use. The development methods studies enabling explanation DNN's has thus blossomed into an active, broad area research. A practitioner...
The rapid expansion of the Internet Things (IoT) and Edge Computing has presented challenges for centralized Machine Deep Learning (ML/DL) methods due to presence distributed data silos that hold sensitive information. To address concerns regarding privacy, collaborative privacy-preserving ML/DL techniques like Federated (FL) have emerged. FL ensures privacy by design, as local participants remains undisclosed during creation a global model. However, performance are insufficient since...
Emerging evidence highlights the role of thyroid hormones in cancer, though findings are controversial. Research on thyroid-related traits lung carcinogenesis is limited. Using UK Biobank data, we conducted bidirectional Mendelian randomization (MR) to assess causal links between cancer risk and dysfunction (hypothyroidism/hyperthyroidism) or function (free thyroxine [FT4], normal-range TSH). Furthermore, smoking-behavior stratified MR analysis, evaluated mediating effect phenotypes...
We introduce a new inference task - Visual Entailment (VE) which differs from traditional Textual (TE) tasks whereby premise is defined by an image, rather than natural language sentence as in TE tasks. A novel dataset SNLI-VE (publicly available at https://github.com/necla-ml/SNLI-VE) proposed for VE based on the Stanford Natural Language Inference corpus and Flickr30k. differentiable architecture called Explainable model (EVE) to tackle problem. EVE several other state-of-the-art visual...
Hurricane Sandy affected the east coast of U.S. in 2012 and posed immense threats to businesses, human lives properties. In order minimize consequent loss a catastrophe like this, critical task disaster management is understand situation updates about from large number disaster-related documents, obtain big picture disaster's trends how it affects different areas. this paper, we present two-layer storyline generation framework which generates an overall or global events first layer, provides...
Current and imminent quantum hardware lacks reliability applicability due to noise limited qubit counts. Quantum circuit cutting — a technique dividing large circuits into smaller subcircuits with sizes appropriate for the resource at hand is used mitigate these problems. However, classical postprocessing involved in generally grows exponentially number of cuts This article introduces notion approximate reconstruction. Using sampling-based method like Markov Chain Monte Carlo (MCMC), we...
Multi-modal learning with both text and images benefits multiple applications, such as attribute extraction for e-commerce products. In this paper, we propose Cross-Modality Attention Contrastive Language-Image Pre-training (CMA-CLIP), a new multi-modal architecture to jointly learn the fine-grained inter-modality relationship. It fuses CLIP sequence-wise attention module modality-wise module. The network uses bridge gap at global level, capture alignment between images. Besides, it...
The ever increasing prevalence of publicly available structured data on the World Wide Web enables new applications in a variety domains. In this paper, we provide conceptual approach that leverages such order to explain input-output behavior trained artificial neural networks. We apply existing Semantic technologies an experimental proof concept.
Map search is a major vertical in all popular engines. It also plays an important role personal assistants on mobile, home or desktop devices. A significant fraction of map traffic comprised “address queries” - queries where either the entire query some terms it refer to address part (road segment, intersection etc.). Here we demonstrate that correctly understanding and tagging are critical for engines fulfill them. We describe several recurrent sequence architectures such queries. compare...
Ki-67 is a key indicator of the proliferation activity tumors. However, no standardized criterion has been established for index calculation. Scale-invariant feature transform (SIFT) algorithm can identify robust invariant features to rotation, translation, scaling and linear intensity changes matching registration in computer vision. Thus, this study aimed develop SIFT-based computer-aided system calculation breast cancer.Hematoxylin eosin (HE)-stained Ki-67-stained slides were scanned...
Abstract Natural products from traditional medicine inherit bioactivity their source herbs. However, the pharmacological mechanism of natural is often unclear and studied insufficiently. Pathway fingerprint similarity based on “drug-target-pathway” heterogeneous network provides new insight into Mechanism Action (MoA) for compared with reference drugs, which are selected approved drugs similar bioactivity. pathway fingerprints may have MoA to drugs. In our study, XYPI, an andrographolide...
A timeline provides a total ordering of events and times, is useful for number natural language understanding tasks. However, qualitative temporal graphs that can be derived directly from text -- such as TimeML annotations usually explicitly reveal only partial orderings times. In this work, we apply prior work on solving point algebra problems to the task extracting timelines annotated texts, develop an exact, end-to-end solution which call TLEX (TimeLine EXtraction). transforms into...
Abstract The role of molecular traits (e.g., gene expression and protein abundance) in the occurrence, development, prognosis lung cancer has been extensively studied. However, biomarkers other layers connections among various that influence risk remain largely underexplored. We conducted first comprehensive assessment associations between (i.e., DNA methylation, expression, metabolite) through epigenome-wide association study (EWAS), transcriptome-wide (TWAS), proteome-wide (PWAS)...
Suzuki coupling reaction of 2-(Trifluormethoxy) phenylboronic acid with 2-bromo-1,3-dichloro-5-nitrobenzene were successfully conducted by using Pd2(dba)3 as the catalyst and in satisfactory to good yields.The 4-(2-(diphenylphosphino)phenyl) morpholine was best ligand this reaction.Biaryl amides 8t which gained much attention pharmaceutical industry because RORγt inhibitors could be synthesized product 3a substrate after 2 steps.Higher yield got method than previous work.
In this paper, we discuss the possibilities of realizing magnetic multivalued (MMV) recording in a coupled multilayer. The hysteresis loop double-layer system is studied analytically, and conditions for achieving MMV are given. from different respects, phase diagrams anisotropic parameters given end.