How does the curse impact neural network training?

It increases parameter space complexity and data requirements.

High-dimensional inputs lead to large weight matrices, increasing overfitting risk and training instability without sufficient regularization.