🤖 AI Summary
Scientific machine learning lacks standardized super-resolution benchmarks tailored to physics-informed data, hindering rigorous method evaluation and real-world deployment. To address this, we introduce SuperBench—the first high-resolution spatiotemporal scientific super-resolution benchmark, covering fluid dynamics, cosmology, and meteorology. Its core contributions are threefold: (1) establishing the first evaluation framework for scientific super-resolution that jointly enforces data fidelity, physical consistency, and degradation robustness; (2) integrating multi-source high-fidelity simulation datasets, complemented by both physics-based consistency verification and data-driven quantitative assessment; and (3) conducting systematic benchmarking of mainstream computer vision models (e.g., EDSR, RCAN) and their physics-guided variants, uncovering pervasive failures—including violations of conservation laws and distortion of turbulent structures. SuperBench provides a unified evaluation platform and establishes baseline performance for physics-informed super-resolution methods.
📝 Abstract
Super-resolution (SR) techniques aim to enhance data resolution, enabling the retrieval of finer details, and improving the overall quality and fidelity of the data representation. There is growing interest in applying SR methods to complex spatiotemporal systems within the Scientific Machine Learning (SciML) community, with the hope of accelerating numerical simulations and/or improving forecasts in weather, climate, and related areas. However, the lack of standardized benchmark datasets for comparing and validating SR methods hinders progress and adoption in SciML. To address this, we introduce SuperBench, the first benchmark dataset featuring high-resolution datasets, including data from fluid flows, cosmology, and weather. Here, we focus on validating spatial SR performance from data-centric and physics-preserved perspectives, as well as assessing robustness to data degradation tasks. While deep learning-based SR methods (developed in the computer vision community) excel on certain tasks, despite relatively limited prior physics information, we identify limitations of these methods in accurately capturing intricate fine-scale features and preserving fundamental physical properties and constraints in scientific data. These shortcomings highlight the importance and subtlety of incorporating domain knowledge into ML models. We anticipate that SuperBench will help to advance SR methods for science.
