论文标题
混合引导程序
The Hybrid Bootstrap
论文作者
论文摘要
找到一种将数值引导与分析LightCone Bootstrap相结合的方法是推进共形性引导程序的重要目标。我们提出了一种混合引导方法来做到这一点。数值和分析性自举方法对光谱的不同区域敏感并相互补充。当它们有效地组合时，混合动力的bootstrap享有两全其美的最好，并且可以显着改善实际CFT的预测。在这项工作中，我们讨论了执行这种混合引导程序的一般策略，并为3D ISING CFT $ \ {σ，ε\} $ System的策略做出了部分实现。即使在相对较低的导数顺序$λ= 19 $下，混合引导程序也可以预测缩放尺寸的非常精确的值$Δ_σ，δ_ε$，该值在前一个$λ= 43 $严格的误差栏内。
Finding a method to combine the numerical bootstrap with the analytic lightcone bootstrap is an important goal to advance the conformal bootstrap program. We propose a hybrid bootstrap method to do just that. The numerical and analytic bootstrap approaches are sensitive to different regions of the spectrum and complement each other. When they are effectively combined, the hybrid bootstrap enjoys the best of both worlds and the prediction for the actual CFT can be significantly improved. In this work, we discuss the general strategy to perform such a hybrid bootstrap, and we make a partial implementation of the strategy for 3D Ising CFT $\{σ,ε\}$ system. Even at relatively low derivative order $Λ=19$, the hybrid bootstrap predicts very precise values for the scaling dimension $Δ_σ,Δ_ε$ that are within the previous $Λ=43$ rigorous error bars.
