Interactive quantum chemistry in virtual reality

Scientists from the Immaterial Reality Laboratory of the University of Bristol (IRL) and ETH Zurich have used virtual reality and artificial intelligence algorithms to know the details of the chemical change.

In a cover story published today in The physical chemistry journal, researchers at the University of Bristol and ETH Zurich describe how advanced interaction and visualization frameworks using virtual reality (VR) enable humans to form algorithms for their work. machine learning and accelerate scientific discoveries.

The team describes its work by designing a state-of-the-art VR open source software framework that can perform quantum mechanics calculations "on the fly".

It allows researchers to explore sophisticated physical models of complex molecular rearrangements involving the creation and breakdown of chemical bonds, the first time that virtual reality has been used to allow such a thing.

The team used its interactive VR system to "teach" quantum chemistry to neural networks.

Lead author, Silvia Amabilino, who works between IRL and Bristol's Center for Computational Chemistry, said: "The generation of datasets to teach quantum chemistry to machines is a long-standing challenge .

"Our results suggest that human intuition, combined with virtual reality, can generate high quality workout data, and thus improve machine learning patterns."

Dr. Lars Bratholm, co-author of the IRL, the Center for Computational Chemistry and the School of Mathematics, added: "For most scientific computing processes, the bottleneck is the processing power But machine learning has created a new bottleneck is the ability to quickly generate high quality data. "

Professor David Glowacki, a researcher at the Royal Society Research and Head of the IRL Department of the Computer Science Department and the Bristol School of Chemistry, said: "Immersive tools like virtual reality offer humans an effective way to express high-level scientific and conceptual knowledge, as we know, this is the first time a virtual reality framework has been used to generate data to drive a neural network. "

The rise of machine learning and automation in science and society has raised important questions about what kind of scientific future we should consciously focus on in the coming decades. Stories of our emerging future often see automation as the ultimate end and it is sometimes difficult to know where the human being fits.

Professor Markus Reiher of ETH added, "This work shows that advanced visualization and interaction frameworks, such as virtual reality and virtual reality, allow humans to complement automated machine learning approaches and accelerate scientific discoveries.

"The document offers an interesting vision of how science can evolve in the near future, where humans focus their efforts on how to train machines effectively."