发表机构
Purdue University; University of Rochester(普渡大学; 罗切斯特大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
针对惯性约束聚变预测中数据稀缺、时间稀疏和峰值敏感的问题,提出首个基于语言模型的ICF-DLM,通过物理分解、双向去噪和物理驱动PPO奖励,在ICFBench上将峰值时间误差从11.6降至9.2步。
AI 中文摘要
惯性约束聚变(ICF)是实现清洁能源的一条主要途径,但在国家点火装置上每次发射的成本约为一百万美元,这使得准确的AI替代模型具有很高的价值。我们研究了外源驱动的ICF波形预测问题,其中必须直接从激光脉冲和目标设计参数推断出512步的中子速率诊断结果,且没有历史响应可观测。该场景对标准时间序列预测器提出了挑战,包括时间稀疏性(纳秒窗口内的皮秒级峰值)、输入输出规模不匹配(少于300次真实发射)以及峰值敏感性(皮秒级时序)。我们提出了ICF-DLM,据我们所知这是首个基于语言模型(LM)的ICF预测器,它结合了:(i) 物理类型化的分解,将波形分解为产额 $Y_{DT}$、峰值时间 $t_{\mathrm{peak}}$ 和局部波形 $w_{\mathrm{local}}$;(ii) 双向去噪,延迟对峰值位置的承诺;(iii) 物理驱动的PPO奖励,将度量结构重新注入数值标记中。在ICFBench(50K模拟+232次实验发射)上,ICF-DLM将峰值时间误差从匹配的自回归LLaMA-3-8B的11.6步降低到9.2步,并优于经典序列模型和基于LLM的时间序列预测器。除ICF外,该方法展示了在数据量低且事件稀疏的科学领域应用的潜力。
英文摘要
Inertial confinement fusion (ICF) is a leading pathway toward clean energy, but each shot at the National Ignition Facility costs on the order of one million dollars, making accurate AI surrogates a high-value target. We study exogenous-driven ICF waveform prediction, where a 512-step neutron-rate diagnostic must be inferred directly from a laser pulse and target design parameters, with no historical response observed. The regime stresses standard time-series predictors with temporal sparsity (picosecond peak in a nanosecond window), input-output scale mismatch (under 300 real shots), and peak sensitivity (picosecond timing). We propose ICF-DLM, to our knowledge the first LM-based ICF predictor, combining (i) a physics-typed decomposition into yield $Y_{DT}$, peak timing $t_{\mathrm{peak}}$, and local waveform $w_{\mathrm{local}}$; (ii) bidirectional denoising that defers commitment to peak location; and (iii) a physics-driven PPO reward re-injecting metric structure across numeric tokens. On ICFBench (50K simulations + 232 experimental shots), ICF-DLM cuts peak-timing error from 11.6 to 9.2 steps over a matched autoregressive LLaMA-3-8B and outperforms classical sequence models and LLM-based time-series predictors. Beyond ICF, the recipe shows potential to address science domains with low data and sparse events.