A team of researchers at the University of Luxembourg, the Technical University of Berlin, and the University of Warwick developed an innovation. The artificial intelligence will be able to predict the electronic properties of molecules and molecular wave functions. It can be used to speed-up the design of new materials or drug molecules.
In the same way, machine learning algorithms and artificial intelligence routinely predict our purchasing behavior. Also, it can recognize our handwriting or faces. Moreover, artificial intelligence is establishing itself as a crucial tool for scientific discovery in scientific research.
Besides, artificial intelligence is widely used in chemistry as well. It can predict simulations of quantum systems or the outcomes of experiments. For achieving this, artificial intelligence needs to incorporate the fundamental laws of physics systematically.
The University of Warwick leads an interdisciplinary team of computer scientists, chemists, and physicists. Thus, the team, alongside the University of Luxembourg and the Technical University of Berlin, developed a deep machine learning algorithm. The algorithm can predict wave functions (states of molecules), which determine all properties of molecules.
Due to learning to solve fundamental equations of quantum mechanics, artificial intelligence achieves this. Nature Communications published that learning happens with the help of unifying quantum chemistry and machine learning with a deep neural network for molecular wavefunctions.
Thus, to solve equations conventionally, it requires massive high-performance computing resources. Consequently, computing time requires months. It is typically the bottleneck to the computational design of new purpose-built molecules for industrial and medical applications. The newly developed artificial intelligence algorithm can supply accurate predictions within seconds on a mobile phone or a laptop.
Artificial Intelligence Progress
Dr. Reinhard Maurer is from the Department of Chemistry at the University of Warwick. He said that it required computer science know-how for developing an algorithm of artificial intelligence. An algorithm flexible enough to capture the behavior and shape of wave functions. It also required physics and chemistry know-how to represent and process quantum chemical data if the form that is manageable for the algorithm. All these required a joint three-year effort.
During an interdisciplinary 3-month fellowship program at IPAM (UCLA), the team has brought together. It was done on the subject of machine learning in quantum physics.
From the Institute of Software Engineering and Theoretical Computer Science at the Technical University of Berlin Professor Dr. Klaus Robert-Muller. He says that interdisciplinary work is significant progress. It shows that artificial intelligence can efficiently perform the most challenging aspects of quantum molecular simulations. Synthetic intelligence methods will establish themselves as an essential part of the discovery process within the next few years. It creates a necessary part of the discovery process in molecular physics and computational chemistry.
From the Department of Physics and Materials Research at the University of Luxembourg is Professor Dr. Alexandre Tkatchenko. He concludes that this work enables a new level of compound design. In this new level, both structural and electronic properties of molecules can tune simultaneously to achieve the desired criteria of applications.
All in all, the team is working hard to incorporate organic materials into quantum mechanics in machine learning and artificial intelligence.
