Physics-Informed Machine Learning Methods For Inverse Design Of Multi-Phase Materials With Targeted Mechanical Properties