AI framework enhances plasma disruption prediction in EAST tokamak

Researchers have developed a novel hierarchical multi-to-single-modal knowledge distillation framework to improve plasma disruption prediction in the EAST tokamak. This method leverages both visible images and time-series diagnostic signals during training to create a robust teacher model. For inference, only the time-series data is used, with knowledge transferred from the teacher to a student model, significantly reducing computational costs while maintaining prediction accuracy. The framework has demonstrated effectiveness on a dataset of 640 EAST discharges. AI

IMPACT This research offers a more efficient method for real-time plasma disruption prediction, crucial for fusion energy research.

RANK_REASON The cluster contains an academic paper detailing a new AI methodology for a specific scientific problem. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• EAST
• Transformer++

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →