A5100410058- Data Visualization and Analytics
- Parallel Computing and Optimization Techniques
- Species Distribution and Climate Change
- Complex Network Analysis Techniques
- Advanced Text Analysis Techniques
- Time Series Analysis and Forecasting
- Cell Image Analysis Techniques
- Aluminum Alloy Microstructure Properties
- Solidification and crystal growth phenomena
- Computer Graphics and Visualization Techniques
- Data Analysis with R
- 3D Shape Modeling and Analysis
- Advanced Neural Network Applications
- Fluid Dynamics and Thin Films
- Algorithms and Data Compression
- Computational Geometry and Mesh Generation
- Biomedical Text Mining and Ontologies
- Advanced Data Storage Technologies
- Aesthetic Perception and Analysis
- Medical Image Segmentation Techniques
- Video Analysis and Summarization
- Multimedia Communication and Technology
- Advanced Vision and Imaging
- scientometrics and bibliometrics research
- Leaf Properties and Growth Measurement
Computer Network Information Center
2013-2024
Chinese Academy of Sciences
2016-2024
Chinese University of Hong Kong
2021
Chinese Academy of Meteorological Sciences
2013
Drexel University
2007-2011
Dimensionality reduction (DR) is a common strategy for visual analysis of labeled high-dimensional data. Low-dimensional representations the data help, instance, to explore class separability and spatial distribution Widely-used unsupervised DR methods like PCA do not aim maximize separation, while supervised LDA often assume certain distributions take perceptual capabilities humans into account. These issues make them ineffective complicated structures. Towards filling this gap, we present...
Time-series data-usually presented in the form of lines-plays an important role many domains such as finance, meteorology, health, and urban informatics. Yet, little has been done to support interactive exploration large-scale time-series data, which requires a clutter-free visual representation with low-latency interactions. In this paper, we contribute novel line-segment-based KD-tree method enable analysis time series. Our enables not only fast queries over series selected regions...
In this paper, we propose SineStream, a new variant of streamgraphs that improves their readability by minimizing sine illusion effects. Such effects reflect the tendency humans to take orthogonal rather than vertical distance between two curves as distance. connect with illusions and doing so provide perceptual foundation for design. As geometry streamgraph is controlled its baseline (the bottom-most curve) ordering layers, re-interpret computation layer algorithms in terms reducing For...
Identifying the significance of specific concepts in diffusion scientific knowledge is a challenging issue concerning many theoretical and practical areas. We introduce an innovative visual analytic approach to integrate microscopic macroscopic perspectives rapidly growing domain. Specifically, our focuses on statistically unexpected phrases extracted from unstructured text titles abstracts at level association with magnitude timeliness their citation impact level. The H-index, originally...
In this paper, we propose the t-FDP model, a force-directed placement method based on novel bounded short-range force (t-force) defined by Student's t-distribution. Our formulation is flexible, exerts limited repulsive forces for nearby nodes and can be adapted separately in its short- long-range effects. Using such graph layouts yields better neighborhood preservation than current methods, while maintaining low stress errors. efficient implementation using Fast Fourier Transform one order...
We present a pyramid-based scatterplot sampling technique to avoid overplotting and enable progressive streaming visualization of large data. Our is based on multiresolution decomposition the underlying density map makes use values in pyramid guide at each scale for preserving relative data densities outliers. show that our competitive quality with state-of-the-art methods runs faster by about an order magnitude. Also, we have adapted it deliver processing chunks updating areas visible...
The emergence of High-Performance Computing (HPC) and Artificial Intelligence (AI) has significantly expanded the applications three-dimensional convolutional neural networks (3D CNNs). At same time, next-generation Sunway supercomputer evidenced its superior computational capabilities in HPC+AI domain. However, complex 3D convolution remains a primary performance limitation many applications. optimization tensor-like operators on processor is usually implemented via multi-level blocking...
User-generated Structured Query Language (SQL) queries are a rich source of information for database analysts, scientists, and the end users databases. In this study group scientists in astronomy computer work together to analyze large volume SQL log data generated by Sloan Digital Sky Survey (SDSS) archive order better understand users' seeking behavior. While statistical analysis such logs is useful at aggregated levels, efficiently exploring specific patterns often challenging task due...
We propose a visualization method to understand the effect of multidimensional projection on local subspaces, using implicit function differentiation. Here, we subspace as neighborhood data points. Existing methods focus points, and information is ignored. Our able analyze shape directional gain more insights into global structure through perception structures. Local subspaces are fitted by ellipses that spanned basis vectors. An accurate efficient vector transformation proposed based...
Addressing the research opportunities we've identified could substantially broaden spectrum of multilingual text-mining and its practicality for supporting global S&T knowledge management. These also share a common set challenges that deserve further attention. For example, competitive intelligence surveillance, which allows organizations to understand their current potential competitors better, often requires extraction names different organizations, technologies, or products from various...
With modern next-generation sequencing (NGS) technology, genome is able to process much more data at low cost. Burrows-Wheeler Aligner (BWA) one of the most widely used open-source software tools align read sequences into a reference and sequence alignment takes time in NGS analysis. BWA-MEM popular BWA. It memory-bound application traditional view cache miss serious problems. But with increase number cores threads on HPC platforms, problem parallelism thread scalability becomes deeply...
Colormapping is an effective and popular visualization technique for analyzing patterns in scalar fields. Scientists usually adjust a default colormap to show hidden by shifting the colors trial-and-error process. To improve efficiency, efforts have been made automate adjustment process based on data properties (e.g., statistical value or histogram distribution). However, as no direct correlation spatial variations, previous methods may be insufficient reveal dynamic range of variations...
We present Target Netgrams as a visualization technique for radial layouts of graphs. Inspired by manually created target sociograms, we propose an annulus-constrained stress model that aims to position nodes onto the annuli between adjacent circles indicating their hierarchy, while maintaining network structure (clusters and neighborhoods) improving readability much possible. This is achieved having more space on than traditional layout techniques. By adapting majorization this model,...
In this paper, we present an integrated framework composed of a highly efficient phase field simulator and in-situ feature extraction library. This novel enables us to conduct extreme-scale micro-structure evolution simulations while the characteristic features each individual grain are extracted on fly. After systematic design optimization new generation Sunway supercomputer, code scales up 39 million cores achieves 582 PFlops in double precision 637 POps mixed precision.
Phase-field models have been used extensively in studying microstructure evolution alloys and the superiority of comprehending, predicting, optimizing microstructure-sensitive macroscopic material properties. The elastic strain energy is a vital factor modeling crystal structure formation solid-solid phase transformations. Conventionally, it computed reciprocal space according to famous Khachaturyan-Shatalov theory. In large-scale simulations, full-space Fourier transform becomes extremely...
Body-fitted curvilinear coordinates are widely used in computational fluid dynamics. In case of covering complex geometry, several such grids may be together with small overlapping region. We present a parallel management tool for equipped Adaptive Mesh Refinement (AMR) capability. By adding fine grid patches or blocks regions the flow where more resolution is needed, as near shocks and detonations, AMR can dramatically speed up computation and/or enable simulations much higher effective...
In this paper, we propose F2-Bubbles, a set overlay visualization technique that addresses overlapping artifacts and supports interactive editing with intelligent suggestions. The core of our method is new, efficient construction algorithm approximates the optimal by considering elements their non-set neighbors. Thanks to efficiency algorithm, achieved, suggestions, users can easily flexibly edit visualizations through direct manipulations local adaptations. A quantitative comparison...
Graph transaction processing raises many unique challenges such as random data access due to the irregularity of graph structures, low throughput and high abort rate relatively large read/write sets in transactions. To address these challenges, we present G-Tran -- an RDMA-enabled distributed in-memory database with serializable snapshot isolation support. First, propose a graph-native store achieve good locality fast for transactional updates queries. Second, adopts fully decentralized...