New dACE method models collective motion with interpretable many-body interactions

Researchers have developed a new method called directional atomic cluster expansion (dACE) to model collective motion, which accounts for many-body interactions unlike previous models that focused on pairwise forces. dACE can accurately identify interaction laws from limited data, outperforming the current competitor, SocialNet, especially in complex scenarios and in 3D. The learned laws are interpretable and have shown better agreement between different data subsets when applied to real-world datasets of fish behavior. AI

IMPACT Introduces a more accurate and interpretable method for modeling complex systems, potentially improving agent-based simulations and behavioral analysis.

RANK_REASON Publication of a new scientific paper detailing a novel methodology. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• Atomic cluster expansion for accurate and transferable interatomic potentials
• directional atomic cluster expansion
• golden shiners

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →