luprox:使用JAX变换的线性不确定性传播

鲁普罗克斯 使用JAX变换的线性不确定性传播 例子 from luprox import linear_uncertainty @ linear_uncertainty def f ( x ): return 1 / ( 1 + jnp . abs ( 0.3 * x ) ** 2 ) # Inputs x = jnp . array ([ 2. ]) covariance = jnp . array ([[ 0.05 ]]) # Covariance matrix for x (scalar for 1D random variables) # Get output and covariance estimates mu_linear , cov_linear = f ( test_point , cov_x ) 结果： STD: 0.04352208226919174 (L