< Back to previous page

Researcher

Attila Kovács

  • Research Expertise:During my Ph.D. research, I studied the physicochemical properties of neoteric reaction media (deep eutectic solvents) and the behavior of enzymes and enzymatic reactions in them. This included computational simulations using various methods. I studied the stability of enzymes and changes in their structure in different solvents using classical molecular dynamics. The properties of the solvents, such as density, viscosity and solubility were estimated by a combined approach of machine learning and group contribution method. In my postdoctoral projects, I'm focusing on the sustainability and recyclability of polymers and polymer-based products. This includes the calculation of the associated environmental impact through life cycle assessment, but also more fundamental simulations. Currently, I'm exploring the application of a hybrid modeling workflow for the separation and substitution of chemically recycled polyurethane streams. In this work, I combine classical and coarse-grained molecular dynamics simulations with conceptual density functional theory and deep neural networks. Some of the projects I am partially involved in include the analysis of previously collected data sets from various chemistry related projects to gain additional understanding of the given problem and to facilitate qualitative predictions.
  • Keywords:MACHINE LEARNING, MOLECULAR DYNAMICS SIMULATIONS, COMPUTATIONAL CHEMISTRY, CHEMINFORMATICS, LIFE CYCLE ASSESSMENT, SUSTAINABILITY, Chemistry (incl. biochemistry)
  • Disciplines:Analytical chemistry, Inorganic chemistry, Macromolecular and materials chemistry, Medicinal and biomolecular chemistry, Organic chemistry, Physical chemistry, Sustainable chemistry, Theoretical and computational chemistry, Other chemical sciences, Biochemistry and metabolism, Other biological sciences, Other natural sciences
  • Research techniques:During my research I became familiar with various computational techniques. I work with small to medium sized datasets to build machine learning solutions, covering the entire workflow from data analysis to the final machine learning models. In my work I mainly focus on regression, as I usually aim for qualitative-structural property relationship models to have predictions on certain physicochemical properties of the studied systems. The other main branch of my technical expertise is molecular level modeling. In the past I mainly worked on classical, all-atom molecular dynamics simulations of enzymes, and recently I am changing my focus to polymer systems and coarse-grained models. I'm also interested in how to combine different modeling techniques to exploit their advantages and mitigate their disadvantages. Recently, I started to extend my knowledge on ab initio methods, including density functional theory calculations. I also have skills in life cycle assessment. Currently, we are providing LCA for a partner company on polymer-based systems.
  • Users of research expertise:Companies, NGOs and other institutions working in the sectors of polymer or polymer-based products formulation or recycling (foams, sealants, paints, adhesives); and interested in the molecular level understanding, prediction or environmental impact of their products and processes.