NEMORA: Neural Equivariant Multipole Operators for Long-Range Atomistic Learning

Key points

• These models describe local atomic environments accurately, but finite spatial cutoffs truncate long-range information flow, and stacking message-passing layers can lead to over-smoothing and over-squashing.1
• Combining learnable long-range equivariant transport with multiscale many-body expressivity and efficient scaling for larger systems remains a central challenge.1
• We introduce Neural Equivariant Multipole Operators (NEMORA), a neural equivariant extension of the Fast Multipole Method (FMM) for learning long-range tensorial representations.1
• NEMORA generalizes the FMM's analytical multipole expansion and translation operators to learned equivariant counterparts on an adaptive spatial hierarchy.1

Sources (1)

• [1]NEMORA: Neural Equivariant Multipole Operators for Long-Range Atomistic Learning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 06:32 PM

  Equivariant graph neural networks have emerged as foundational architectures for machine-learned interatomic potentials, approaching quantum-chemical accuracy at a fraction of the computational cost.

  These models describe local atomic environments accurately, but finite spatial cutoffs truncate long-range information flow, and stacking message-passing layers can lead to over-smoothing and over-squashing.

Extractive summary: sentences quoted from the sources.