The virtual fields method (VFM) is an inverse mechanics technique leveraging full-field deformation data and equilibrium principles. Focusing on hyperelastic constitutive models for soft biological tissues, a significant number of recent advances have contributed to the popularization of the VFM. In the talk, we will review: 1. recent formulations of a more accessible and modern VFM where virtual fields are automatically generated, 2. enhancement of robustness for the identification of regionally variable constitutive parameters, 3. machine-learning integration in the VFM spanning from the formulation of surrogate constitutive equations with artificial neural networks to the resolution of partial differential equations. We will show how these recent advances pave the way to real-time clinical applications of the VFM for complex soft tissue modeling.