Federal Project Grant Award Summary

Oregon State University received a $299,354 Project Grant from the National Science Foundation's Office of Advanced Cyberinfrastructure under the Computer and Information Science and Engineering program (CFDA 47.070) for the collaborative research project "Mitigating Artifacts in Scientific Data Compressors with a Learning-Driven Framework." The award, effective July 1, 2026, through September 30, 2028, supports the development of a novel deep learning framework designed to identify and mitigate compression artifacts produced by error-controlled lossy compressors used in scientific applications. The deliverables include comprehensive characterization of compression artifacts across raw data and post-hoc analyses, deep learning models optimized through transfer learning to recover data quality while minimizing training costs, output fusion mechanisms to preserve features of interest, and uncertainty quantification methods for validating recovered data quality.

The framework will be integrated into state-of-the-art lossy compression systems and validated across multiple scientific domains including climatology, cosmology, fusion energy science, and X-ray ptychography. By enabling more effective use of existing compression technologies for data storage, transmission, and analytics, the project aims to improve data integrity in scientific research while reducing computational and storage costs. The work advances cyberinfrastructure capabilities in data management and contributes to research and education initiatives in advanced computing infrastructure across multiple scientific disciplines.