用低成本模拟提升高精度参数估计，大幅降低计算成本。
Transfer learning for multifidelity simulation-based inference in cosmology
- 利用低精度模拟预训练，再结合少量高精度模拟进行迁移学习。
- 所需高精度模拟数量减少8到15倍，效果依赖模型复杂度和维度。
- 适合需要高效宇宙学参数推断的研究者，尤其关注计算资源受限场景。
基于模拟的推断（SBI）在缺乏显式似然函数或模型时，可用于宇宙学参数估计。然而，SBI依赖机器学习进行神经压缩与密度估计，需大量训练数据，而高质量模拟代价高昂。本文提出多保真度迁移学习方法，结合低成本低精度模拟与有限高精度模拟，解决此问题。我们在两个独立模拟套件的暗物质密度图上验证该方法，基于仅含暗物质的N体模拟预训练，使高精度流体动力学模拟需求量减少8至15倍，具体取决于模型复杂度、后验维度及评估指标。该方法在显著降低计算成本的同时，仍实现高性能与高精度推断。
原文摘要 · Abstract (English)
Simulation-based inference (SBI) enables cosmological parameter estimation when closed-form likelihoods or models are unavailable. However, SBI relies on machine learning for neural compression and density estimation. This requires large training datasets which are prohibitively expensive for high-quality simulations. We overcome this limitation with multifidelity transfer learning, combining less expensive, lower-fidelity simulations with a limited number of high-fidelity simulations. We demonstrate our methodology on dark matter density maps from two separate simulation suites in the hydrodynamical CAMELS Multifield Dataset. Pre-training on dark-matter-only $N$-body simulations reduces the required number of high-fidelity hydrodynamical simulations by a factor between $8$ and $15$, depending on the model complexity, posterior dimensionality, and performance metrics used. By leveraging cheaper simulations, our approach enables performant and accurate inference on high-fidelity models while substantially reducing computational costs.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。