Using deep learning to solve fundamental problems in computational quantum physics and chemistry from first principles.
Quantum mechanics is the fundamental theory of nature that describes the behavior of matter and energy at atomic and subatomic scales. It is the foundation of chemistry and materials science, and underpins our understanding of the universe. But solving the equations of quantum mechanics for systems of more than a few particles is a computational challenge that grows exponentially with system size. This has limited our ability to accurately predict the properties of molecules and materials from first principles.
FermiNet is a new deep learning architecture that can predict the quantum mechanical properties of molecules and materials with high accuracy, using only the fundamental laws of physics as input. By combining the expressive power of deep neural networks with the principles of quantum mechanics, FermiNet can predict the electronic structure of molecules and materials with unprecedented accuracy and efficiency. This opens up new possibilities for understanding and designing molecules and materials for a wide range of applications, from drug discovery to renewable energy.
We have applied FermiNet to a wide range of problems in quantum chemistry and condensed matter physics, and have demonstrated its ability to predict the properties of molecules and materials with high accuracy. Our results show that FermiNet can achieve state-of-the-art accuracy on a wide range of benchmark datasets, and can scale to large systems with thousands of atoms. This opens up new possibilities for using deep learning to solve fundamental problems in computational quantum physics and chemistry, and paves the way for a new era of discovery and innovation.