The goal of this project is to identify fundamentally validated mean-field and full-field models capable of predicting failure in Mg alloys. This is a critical step for developing materials design concepts to enable the use of lightweight Mg alloys in safety critical applications or development of deformation processing strategies. A critical, fundamental gap in the understanding of Mg alloy deformation relates to dislocation-dislocation, dislocation-twin, twin-twin, and twin-grain boundary (GB) interactions and their effects on strain hardening and damage initiation. A multi-scale approach is required, because key mechanisms operate at different length scales: 1. atomistic (dislocation core interactions with other dislocations and with twin boundaries, and decohesion); 2. microscopic (dislocation-dislocation, dislocation-twin, and twin-twin interactions); 3. mesoscopic (twin-parent grain and grain-grain compatibility interactions leading to backstress); and 4. macroscopic (applications to forming simulation or performance prediction design incorporating shear localization). These models will greatly aid efforts to render lightweight Mg alloys "formable" and "crushable," so that society can exploit performance and efficiency benefits.