研究人员开发了“The Neural Compiler”，一个将符号程序转换为可微分 PyTorch 模块的系统，用于科学机器学习。这种方法可以精确地将已知的物理知识编码到混合模型中，而学习到的组件则处理未知方面。该编译器在恢复物理常数和处理复杂方程链方面表现出高精度和可组合性，显著优于标准的物理信息神经网络 (PINNs)。 AI
影响 通过将符号物理与神经网络相结合，实现了更准确、更具可组合性的科学机器学习模型。
排序理由 该集群包含一篇描述新系统及其评估的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2605.22498v1 Announce Type: new Abstract: Scientific machine learning often requires combining known physics with unknown parameters or correction terms learned from data. Existing approaches either ignore known structure, encode it as a soft penalty, or require hand-writte…
Scientific machine learning often requires combining known physics with unknown parameters or correction terms learned from data. Existing approaches either ignore known structure, encode it as a soft penalty, or require hand-written PyTorch code for each equation. We present The…