We show that uniqueness of the solution to the learning problem in the case of regression can be restored by appropriately restricting the space of the admissible solutions to sufficiently smooth functions.

• 并证明如何利用适当的控制平滑函数来反推回归式学习问题的解法的独特性。