MIT lab's PEM-UDE recovers interpretable chaos equations from noisy data
burny_tech · x · 2026-09-29
Researchers at MIT's Miller Lab and Christopher Rackauckas et al. present PEM-UDE (arXiv 2507.03631), combining prediction-error methodology with universal differential equations to discover governing equations from limited, noisy chaotic data. Key points:
- Prediction-error feedback smooths chaotic optimization; noise and model misspecification introduce a gain-dependent stability-bias trade-off.
- On the Rossler attractor and a real circuit, it recovers correct functional forms even with noise 5x the signal magnitude in one observed dimension.
- The method accepts prior knowledge as an initial functional form, used to learn neural circuit equations with sparse connectivity that conventional neural mass models lack.
- The authors note zero-loss-set preservation doesn't guarantee unique structural identifiability, and the approach yields testable predictions about neural populations.
More from Research
- Apollo Research: Models in Coding Evals Favor Graders Over Users, Reward-Seeking Grows With RL — burny_tech · 2026-09-29
- Late-layer neurons in Qwen act like on-off switches, unlike Olmo — Sauers_ · 2026-09-29
- PyroAdapt lifts California wildfire prediction, boosting extreme-fire recall by 39.7 points — UIUC-CS · 2026-09-29
- KAIST proposes EAPO: entropy-guided credit assignment reinforcing surprising success in RLVR — kaist-ai · 2026-09-29
- Perturbed public documents synthesize training data, lifting Qwen 35B to trillion-param level — AllSpark-Research · 2026-09-29
- Skill2Env synthesizes executable environments from skills for agent post-training — AllSpark-Research · 2026-09-29