Past Event: Oden Institute Seminar

Hierarchical Autoencoder-based Lossy Compression for Large-scale High-resolution Scientific Data

Jian Tao, Texas A&M University

Abstract

Lossy compression has become an important technique to reduce data size in many domains. This type of compression is especially valuable for large-scale scientific data, whose size ranges up to several petabytes. Although Autoencoder-based models have been successfully leveraged to compress images and videos, such neural networks have not widely gained attention in the scientific data domain. Our work presents a neural network that not only significantly compresses large-scale scientific data but also maintains high reconstruction quality. The proposed model is tested with scientific benchmark data available publicly and applied to a large-scale high-resolution climate modeling data set. Our model achieves a compression ratio of 140 on several benchmark data sets without compromising the reconstruction quality. Simulation data from the High-Resolution Community Earth System Model (CESM) Version 1.3 over 500 years are also being compressed with a compression ratio of 200 while the reconstruction error is negligible for scientific analysis.

Biography

Dr. Jian Tao is an Assistant Professor from the Section of Visual Computing & Computational Media in the School of Performance, Visualization & Fine Arts at Texas A&M University. He is also the Director of the Digital Twin Lab and the Assistant Director for Project Development at the Texas A&M Institute of Data Science. Tao received his Ph.D. in Computational Astrophysics from Washington University in St. Louis in 2008 and worked on computational frameworks for numerical relativity, computational fluid dynamics, coastal modeling, and other applications at Louisiana State University before he joined Texas A&M in 2016. In 2018, Tao led the Texas A&M team to the final of both the ASC18 and SC18 student cluster competitions. He is a faculty advisor of the Texas A&M 12th Unmanned Team for the SAE/GM AutoDrive Challenge Competition. Supported by a grant from the Department of Commerce, Tao is leading an effort to build a digital twin for the Disaster City managed by the Texas A&M Engineering Extension Service. Tao is an NVIDIA DLI University Ambassador and a contributor to the SPEC CPU 2017 benchmark suite. He currently serves as the Testbed Committee Co-Chair of the IEEE Public Safety Technology Initiative. His research interests include digital twin, numerical modeling, machine learning, data analytics, distributed computing, visualization, and workflow management.