• Nathaniel Gorski, Xin Liang, Hanqi Guo, Bei Wang. FeatureZ: A General Framework for Feature-Preserving Compression via Pointwise Bounds and Star Classification. Accepted at IEEE Transactions on Visualization and Computer Graphics. 2027.

arxiv github

• Kuangshi Ai, Haichao Miao, Kaiyuan Tang, Nathaniel Gorski, Jianxin Sun, Guoxi Liu, et al. SciVisAgentBench: A Benchmark for Evaluating Scientific Data Analysis and Visualization Agents. Accepted at IEEE Transactions on Visualization and Computer Graphics. 2027.

arxiv project webpage

• Logan Bolton, Tomas Karella, Nathaniel Gorski, Emily Shinkle, Alice E. A. Allen, Pieter J Swar, et al. Rotation Invariant Machine Learning for Highly Symmetric Molecular Configurations. Accepted at NeurIPS 2026 Workshop on Symmetry and Geometry in Neural Representations (Extended Abstracts Track). 2026.

• Alex Fallin, Nathaniel Gorski, Tripti Agarwal, Bei Wang, Ganesh Gopalakrishnan, Martin Burtscher. Fast Topology-Aware Lossy Data Compression with Full Preservation of Critical Points and Local Order. IEEE Transactions on Big Data. 2026.

arxiv github

• Weiran Lyu, Nathaniel Gorski, Jeff Phillips, Bei Wang. A Topology-Preserving Coreset for Kernel Regression in Scientific Visualization. IEEE Transactions on Visualization and Computer Graphics. 2026. Best Paper Award IEEE PacificVis 2026 (TVCG journal track).

pdf

• Nathaniel Gorski, Xin Liang, Hanqi Guo, and Bei Wang. TFZ: Topology Preserving Compression of 2D Symmetric and Asymmetric Tensor Fields. IEEE Transactions on Visualization and Computer Graphics. 2026.

arxiv github

• Nathaniel Gorski, Xin Liang, Hanqi Guo, Lin Yan, and Bei Wang. A general framework for augmenting lossy compressors with topological guarantees. IEEE Transactions on Visualization and Computer Grpahics. 2025.

arxiv github

Topology-Preserving Compression of 2D Tensor Fields

• Designed a framework that augments any error-bounded lossy compressor to preserve topological features of 2D symmetric and asymmetric tensor fields while maintaining a user-specified error bound

• Wrote custom scientific visualizations for tensor fields

Extracting the Topology of Loss Landscapes

•  Currently studying how the topology of the loss function of a deep neural network interacts with training

• Visualized the loss function of deep neural networks on 2D linear subspaces of the parameter space

Contour Tree Preserving Data Compression

• Designed a framework that augments any lossy compressor to preserve the contour tree of scalar field data which is applicable to both classic and deep learning data compressors

• Compared to the state-of-the-art, improved compression ratios by 2.8-6.8x and reduced running times by 25-95% while maintaining similar reconstruction quality

I wrote an undergraduate honors thesis in order to receive my BS with honors. It is a summary of an independent study that I completed on differential topology. None of the material was copied verbatum, and many of the proofs are completely original.

My undergraduate thesis was titled "Oriented Intersection in Differential Topology". My thesis advisor was Craig Westerland. You can read it here.

To receive my M.A.T., I wrote an extended literature review about perfectionism among adolescents. The title of the paper is "The Causes and Effects of Perfectionism Among High School Students." My advisor was Il-Hee Kim. You can read it here.

I did research in algebraic combinatorics during Spring 2018 with one of my graduate TAs. The problem that I sought to solve was to find the largest subset of a cyclic group of order n such that no k elements of the subset summed up to 0, for any given n and k. I solved the problem in the case where n and k meet certain conditions.

This research was inspired by the work of Bela Bajnok. I worked under Ryan Matzke, a former student of Bajnok.

I presented the results of my research at the CeSMR Conference at the University of Nebraska - Lincoln.

You can find the writeup of my research here. These notes were written to summarize what I had done, so this does not represent my best technical writing.