Thrilled to announce our new paper diving into Zeus' decentralized mixture-of-experts approach for machine learning-based weather prediction (MLWP)! ⚡️ To assess the subnet’s optimal capability, we introduced the Best Recent Performer (BRP) metric, which identifies the top miners for each forecast horizon. Notably, the BRP achieves an average RMSE reduction of 39.8% on 2m temperature, while the collective mean of all subnet miners improves upon the baseline by 33.7%. Read full paper👇 🔗