NeurIPS 2026 Competition: Predicting Fusion Plasma Equilibrium with ML
wattmaller1 · x · 2026-08-05
Sophelio has launched the "Fusion Equilibrium Challenge" for the NeurIPS 2026 Competition Track, featuring a $1,000 prize pool. The competition explores using machine learning models to predict the geometry and key equilibrium properties of fusion plasma without relying on magnetic sensor inputs.
Open Dataset
- Contains 9,121 experimental plasma shots from the DIII-D and MAST tokamaks.
- Features 98 GB of data, providing poloidal-field coil currents and plasma current.
Task
Participants must predict a full (65×65) poloidal-flux map and derive plasma boundaries. The challenge aims to solve engineering issues where intense neutron radiation degrades magnetic sensors in future reactor environments.
More from Research
- AI-Generated Stories Beat Human Ones for Readability, Study Finds — nordicinst · 2026-08-05
- Researcher: No Downside to Publishing Sloppy Papers If You Outrun the Blast — RylanSchaeffer · 2026-08-05
- New Insight: Interpreting Bregman Divergences as a Weighted Power Distance — FrnkNlsn · 2026-08-05
- Netflix Introduces GenRec: An LLM-Native Recommendation System Outperforming Traditional Models — rseroter · 2026-08-05
- Reverse-Engineering NVIDIA Blackwell Tensor Cores for Bit-for-Bit Software Simulation — ycombinator · 2026-08-05
- Analyzing the Four Core Paradigms of Modern In-Context TTS — rdesh26 · 2026-08-05