Self-Scalable Tanh (Stan): Multi-Scale Solutions for Physics-Informed Neural Networks | IEEE Journals & Magazine | IEEE Xplore