(4be) Computational Study of Materials Structures and Phase Transitions for Energy and Healthcare Applications Using Molecular Dynamics and Machine Learning Algorithms

Area 1: Molecular Dynamics simulations of phase transitions and Material structure analysis:

Manufacturing of new materials with different physical properties requires an understanding of the related phase transitions. Indeed, the structure of a material is strongly related to the conditions under which the phase transition is occurring. Therefore, using Molecular Dynamics simulations, we investigated different phase transitions for different types of systems such as metallic and molecular (water) systems. We showed the impact of the thermodynamic conditions on the structure of the materials formed and the emerging polymorphisms/polyamorphisms. This implies navigating the phase diagram of each compound to explore the impact of both degrees of supercooling and pressure on solid-liquid (nucleation and crystallization) and liquid-liquid phase transition. The understanding of crystal and amorphous structure formations is of high relevance for both energy and medical applications in the development of materials with tailored properties.

Area 2: Cancer cell detection and prediction through deep learning Algorithms:

Cancer-related diseases are one of the largest causes of death in our current society. However, their early detection leads to higher chances of recovery. Therefore, we have developed deep learning models for the early detection of cancerous cells. Because a high degree of accuracy is required for proper diagnosis, a comparison of common metrics between different algorithms (Classification Tree, Naives Bayes, etc.) was performed. Deep learning algorithms may be applied to various fields increasing the performance of classification or prediction tasks.