Simulation vs. Understanding: A Tension, in Quantum Chemistry and Beyond. Part A. Stage Setting.
Roald HoffmannJean-Paul MalrieuPublished in: Angewandte Chemie (International ed. in English) (2020)
We begin our tripartite Essay with a triangle of understanding, theory and simulation. Sketching the intimate tie between explanation and teaching, we also point to the emotional impact of understanding. As we trace the development of theory in chemistry, Dirac's characterization of what is known and what is needed for theoretical chemistry comes up, as does the role of prediction, and Thom's phrase "To predict is not to explain." We give a typology of models, and then describe, no doubt inadequately, machine learning and neural networks. In the second part, we leave philosophy, beginning by describing Roald's being beaten by simulation. This leads us to artificial intelligence (AI), Searle's Chinese room, and Strevens' account of what a go-playing program knows. Back to our terrain-we ask "Quantum Chemistry, † ca. 2020?" Then move to examples of AI affecting social matters, ranging from trivial to scary. We argue that moral decisions are hardly to be left to a computer. At this point, we try to pull the reader up, giving the opposing view of an optimistic, limitless future a voice. But we don't do justice to that view-how could we? We return to questioning the ascetic dimension of scientists, their romance with black boxes. Onward: In the 3rd part of this Essay, we work our way up from pessimism. We trace (another triangle!) the special interests of experimentalists, who want the theory we love, and reliable numbers as well. We detail in our own science instances where theory gave us real joy. Two more examples-on magnetic coupling in inorganic diradicals, and the way to think about alkali metal halides, show us the way to integrate simulation with theory. Back and forth is how it should be-between painfully-obtained, intriguing numbers, begging for interpretation, in turn requiring new concepts, new models, new theoretically grounded tools of computation. Through such iterations understanding is formed. As our tripartite Essay ends, we outline a future of consilience, with a role both for fact-seekers, and searchers for understanding. Chemistry's streak of creation provides in that conjoined future a passage to art and to perceiving, as we argue we must, the sacred in science.