Benchlab is FDL’s IniTiative to maintain SOTA in ML FOR science.

Coming soon

This initiative aims to bridge the gap between model development and operational deployment by challenging participants to predict ambient solar wind speed three days in advance.

While recent advances in AI have enabled new predictive methods, operationalizing models remains a challenge. This tournament aims to test submitted workflows in increasingly realistic operational scenarios.

The primary goal is to improve critical space predictive capabilities by using the community to maintain efficacy. Submissions will be tested against real-world, real-time data to benchmark performance against physics-based baselines.

The competition is structured into three iterative "Challenge Laps," designed to simulate the demands of an operational environment. Each lap becomes progressively more demanding, with code updates permitted between phases.