This symposium is inspired by the Materials Genome Initiative and is focused on study of synthetic and natural soft-materials using theoretical and computational modeling, and machine-learning methods. The nature of nanoscopic units (e.g. biomolecules, polymers etc.) and their macroscopic structures (e.g. micelles, fibers, bilayers, vesicles, helices, emulsions, foams etc.) often determine the properties and applications of the architectures of soft-materials. Recent advances in the experimental characterization techniques have been helpful in improving our understanding of these systems to a certain extent. However, it is still very challenging to predict the final macroscopic structure, properties, stimuli-sensitivity, and emerging collective phenomena, a priori, from knowledge of the atomic constituents of nanoscopic units and processing parameters (solvent, temperature, gas environments etc.). This definitely limits our ability to further develop and improve new soft materials with predefined structure, properties, and function. To overcome these limitations, to test and validate the design, and to predict the characteristics of architectures of soft-materials with precision, advanced multiscale modeling methods and theory combined with machine learning and the state-of-art high performance computing has been employed. This symposium will discuss the design and characterization of structural and dynamical properties of soft materials using machine learning and computational modeling.