Advanced and frontier

Scaling attention to millions of points, joint-embedding prediction, how the main model lineages evolved, and recent papers.

• Scaling to millions of points: Attention grows with the square of the points; latent and linear schemes keep it tractable.
• JEPA: Predict the latent of the flow from geometry and condition; decode a field only when you need one.
• Lineages: Four families of field surrogates, from their founding idea to 2026.
• Recent papers: Recent field-surrogate architectures from NeurIPS, ICML and ICLR, 2023 to 2026, in order.