A5076622540- Data Visualization and Analytics
- Climate variability and models
- Meteorological Phenomena and Simulations
- Computer Graphics and Visualization Techniques
- Data Analysis with R
- Species Distribution and Climate Change
- Electron and X-Ray Spectroscopy Techniques
- Scientific Computing and Data Management
- Atmospheric and Environmental Gas Dynamics
- Fault Detection and Control Systems
- Machine Learning in Materials Science
- Advanced Chemical Sensor Technologies
- Nuclear reactor physics and engineering
- Environmental and Agricultural Sciences
- Remote Sensing and Land Use
- Solid-state spectroscopy and crystallography
- Explainable Artificial Intelligence (XAI)
- Remote Sensing in Agriculture
- Advanced Clustering Algorithms Research
- Video Analysis and Summarization
- Generative Adversarial Networks and Image Synthesis
- Adversarial Robustness in Machine Learning
- Computational Geometry and Mesh Generation
- Complex Network Analysis Techniques
- Time Series Analysis and Forecasting
SRI International
2024-2025
Los Alamos National Laboratory
2020-2024
Fujitsu (United States)
2024
Palo Alto Research Center
2023-2024
Menlo School
2024
Arizona State University
2022
The Ohio State University
2015-2020
National Institute of Technology Durgapur
2012
Over the last decade, ensemble visualization has witnessed a significant development due to wide availability of data, and increasing needs from variety disciplines. From data analysis point view, it can be observed that many works focus on same facet use similar aggregation or uncertainty modeling methods. However, lack reflections those essential commonalities systematic overview prevents researchers effectively identifying new unsolved problems planning for further developments. In this...
Building on top of the success in AI-based atmospheric emulation, we propose an ocean emulation and downscaling framework focusing high-resolution regional over Gulf Mexico. Regional presents unique challenges owing to complex bathymetry lateral boundary conditions as well from fundamental biases deep learning-based frameworks, such instability hallucinations. In this paper, develop a autoregressively integrate ocean-surface variables Mexico at $8$ Km spatial resolution without unphysical...
Abstract Due to their high thermal efficiency and long functional life, diesel engines have become ubiquitous in automobiles. Diesel are vulnerable component failure sensor faults. New cognitive fault diagnosis algorithms crucial for the safe operation of equipment. Conventional model-based approaches limited capabilities owing approximations made during development these models. In comparison, efficacy most data-driven depends on quantity data. Additionally, existing do not consider...
Distributions are often used to model uncertainty in many scientific datasets. To preserve the correlation among spatially sampled grid locations dataset, various standard multivariate distribution models have been proposed visualization literature. These treat each location as a univariate random variable which at that location. Standard distributions (both parametric and nonparametric) assume all marginals of same type/family distribution. But reality, different show statistical behavior...
Complex computational models are often designed to simulate real-world physical phenomena in many scientific disciplines. However, these simulation tend be computationally very expensive and involve a large number of input parameters which need analyzed properly calibrated before the can applied for real studies. We propose visual analysis system facilitate interactive exploratory high-dimensional parameter space complex yeast cell polarization simulation. The proposed assist biologists, who...
Patients managing a complex illness such as cancer face information challenge where they not only must learn about their but also how to manage it. Close interaction with healthcare experts (radiologists, oncologists) can improve patient learning and thereby, disease outcome. However, this approach is resource intensive takes expert time away from other critical tasks. Given the recent advancements in Generative AI models aimed at improving system, our work investigates whether generative...
CoDDA (Copula-based Distribution Driven Analysis) is a flexible framework for large-scale multivariate datasets. A common strategy to deal with scientific simulation data partition the domain and create statistical summaries. Instead of storing high-resolution raw from simulation, compact summaries results in reduced storage overhead alleviated I/O bottleneck. Such summaries, often represented form probability distributions, can serve various post-hoc analysis visualization tasks. However,...
Visualizing the similarities and differences among an ensemble of isosurfaces is a challenging problem mainly because cannot be displayed together at same time. For isosurfaces, visualizing these spatial surfaces essential to get useful insights as how individual simulations affect different isosurfaces. We propose scheme visualize variations with respect statistically significant within ensemble. Understanding such regions helpful in analyzing influence runs over domain. In this regard, we...
Convolutional neural networks are one of the most important and widely used constructs in natural language processing AI general. In many applications, they have achieved state-of-the-art performance, with training time faster than other alternatives. However, due to their limited interpretability, less favored by practitioners over attention-based models, like RNNs self-attention (Transformers), which can be visualized interpreted more intuitively analyzing attention-weight heat-maps. this...
Recent advances in the synthesis of polar molecular materials have produced practical alternatives to ferroelectric ceramics, opening up exciting new avenues for their incorporation into modern electronic devices. However, order realize full potential polymer and crystals technological applications, it is paramount assemble evaluate all available data such compounds, identifying descriptors that could be associated with an emergence ferroelectricity. In this paper, we utilized data-driven...
Uncertainty of scalar values in an ensemble dataset is often represented by the collection their corresponding isocontours. Various techniques such as contour-boxplot, contour variability plot, glyphs and probabilistic marching-cubes have been proposed to analyze visualize All these assume that a value interest already known user. Not much work has done guiding users select for uncertainty analysis. Moreover, analyzing visualizing large isocontours selected its own challenges. Interpreting...
We study the sensitivity of South Asian Summer Monsoon (SASM) precipitation to Southern Hemisphere (SH) subtropical Absorbed Solar Radiation (ASR) changes using Community Earth System Model 2 simulations. Reducing positive ASR biases over SH subtropics impacts SASM, and is sensitive ocean basin where are imposed. Indian Ocean (IO) shifts rainfall equatorial IO northward causing 1-2 mm/day drying south equator, Pacific increases northern continental regions by mm/day, Atlantic have little...
Abstract We study the sensitivity of South Asian Summer Monsoon (SASM) precipitation to Southern Hemisphere (SH) subtropical Absorbed Solar Radiation (ASR) changes using Community Earth System Model 2 simulations. Reducing positive ASR biases over SH subtropics impacts SASM, and is sensitive ocean basin where are imposed. Indian Ocean (IO) shifts rainfall equatorial IO northward causing 1–2 mm/day drying south equator, Pacific increases northern continental regions by mm/day, Atlantic have...
A significant challenge on an exascale computer is the speed at which we compute results exceeds by many orders of magnitude save these results. Therefore Exascale Computing Project (ECP) ALPINE project focuses providing exascale-ready visualization solutions including in situ processing. In and analysis runs as simulation run, simulations are they generated avoiding need to entire storage for later analysis. The made post hoc tools, ParaView VisIt, ready developed algorithms...
Data modeling and reduction for in situ is important. Feature-driven methods data analysis are a priority future exascale machines as there currently very few such methods. We investigate deep-learning based workflow that targets processing using autoencoders. propose Residual Autoencoder integrated Dense Block (RRDB) to obtain better performance. Our proposed framework compressed our test into 66 KB from 2.1 MB per 3D volume timestep.
We provide a visualization based answer to understanding the evolution and structure of dark matter halos by addressing tasks assigned in 2015 IEEE Scientific Visualization Contest. The data released this year is Cosmological Simulation dataset generated from Dark Sky experiments. Out we are following: integration browsing, halo identification diving deep into substructure.
In this study, we propose a solution to estimating system responses external forcings or perturbations. We utilize the Fluctuation-Dissipation Theorem (FDT) from statistical physics extract knowledge using an AI model that can rapidly produce scenarios for different by leveraging FDT and analyzing large dataset Earth System Models. Our model, AiBEDO, accurately captures complex effects of radiation perturbations on global regional surface climate, enabling faster exploration impacts...
Clouds play a vital role both in modulating Earth's radiation budget and shaping the coupled circulation of atmosphere ocean, driving regional changes temperature precipitation. However, climate response to changing clouds is one largest uncertainties Earth System Models (ESM) when producing decadal (and longer) projections. This limitation becomes apparent especially analyzing scenarios with large cloud properties, e.g., presence greenhouse gases (leading loss or engineered interventions...