Based on the partnership between the University of South Alabama (USA) and DOE National Lab, this research will establish overall solutions for neuromorphic systems with robust immunity to non-ideal properties. In particular, using a DOE-funded user facility at Center for Integrated Nanotechnologies (CINT)and collaborating with DOE scientists, memrisor arrays will be fabricated and an optimized mask will be developed to maximumly avoid sneak path current (SPC) and heighten the performance of neuromorphic systems for both on-line and off-line learning scenarios. This research uniquely connects machine learning (computer science), neuromorphic hardware (electrical engineering, computer engineering, and neurological engineering), and semiconductor manufacture (mechanical engineering and chemical engineering), thereby creating new research directions for multidisciplinary study and significantly enhancing multidisciplinary workforce development.