Deepmind Strikes Back, Now Tackling Quantum Mechanical Calculations

Alphabet's Deepmind's success on biological chemistry with AlphaFold is by now well established as I presented and discussed in previous stories, paving a new era of biological chemistry, biophysics, and well… biology in general. Now, the AI company just struck back, this time tackling so-called "quantum mechanical (QM) calculations", which deal with the highest possible "resolution" when studying chemistry: electron densities, distributions, and spin states--the key elements modulating reactivity and bulk properties. It turns out that the structure and reactivity of molecules (small, big, or "infinite" as in blocks of materials; organic, inorganic or mixed) is determined by their electronic structures, which of course depend on 3D structures but are not just about atom distribution, rather more about electron distribution. To be more precise, it's not even just about how electrons distribute, but also relates to their spin states (i.e. These "electronic structures" of molecules can globally be described by the Schrödinger equation, through so-called "electron wavefunctions" which are essentially mathematical descriptions of the probability of finding an electron in a particular location in space.