JEPA controller for PDEs cuts tracking error 53% on out-of-distribution targets
burny_tech · x · 2026-07-28
- A paper proposes a goal-agnostic JEPA for closed-loop control of PDEs, focusing on the Navier–Stokes system.
- It uses a kinetic-energy probe to define control objectives instead of directly minimizing latent-space L2 distance, which can distort physical error signals.
- The authors report 53% average improvement in physical state tracking on out-of-distribution aperiodic targets, and 2.66% mean relative error when regulating kinetic energy.
More from Models
- Artificial Analysis Releases 2025 Year-End State of AI and Trends Report — ArtificialAnlys · 2026-07-28
- Kimi K3 narrows the open-weights gap to just 4 points on Artificial Analysis — ArtificialAnlys · 2026-07-28
- inclusionAI’s LLaDA2.2-flash is trending on Hugging Face — inclusionAI · 2026-07-28
- Indie PPO experiment reaches 95% accuracy in 3–4 minutes on about $20 — ctjlewis · 2026-07-28
- User says they are switching back to Opus 4.8 because the 5.0 series is hard to steer — Aizkmusic · 2026-07-28
- Analyst ranks Opus 5 third overall but calls it the best coding model — JasonBotterill · 2026-07-28