Accelerated chemical science with AI

Abstract

The ASLLA Symposium focused on accelerating chemical science with AI. Discussions on data, new applications, algorithms, and education were summarized. Recommendations for researchers, educators, and academic bodies were provided.

Theoretical Design Strategies for Area-Selective Atomic Layer Deposition

“…Recently, researchers are actively developing automated transition state search algorithms, , and their application to surface and solid state reactions is anticipated. Also, machine-learning (ML) or artificial intelligence (AI) techniques are emergent in chemical materials science. − An application of ML toward surface chemical research is the generation of interatomic potentials or force fields with chemical level accuracy. − This approach utilizes machine learning to generate force fields based on large amounts of data set, often gathered using DFT calculations, aiming to provide results that are more accurate than those of conventional force fields. In particular, universal interatomic potential that covers most of the periodic table can facilitate large-scale simulations of chemical reactions involving diverse elements in large scales, eliminating the need for case-by-case optimization of the force fields. − , …”

Comment on “Comparing the Performance of College Chemistry Students with ChatGPT for Calculations Involving Acids and Bases”

“…Alternatively, the ability to generate reliable , human-readable explanations on demand means AI is more generally applicable than Clark et al suggest. Third, there is little doubt that AI in general, , and LLMs in particular, , provide a powerful new tool for chemistry, so teaching its effective use is imperative. The procedure described hereproviding tools and structured reasoning promptsis already being used by ChemCrow and AI-Coscientist to direct chemical research.…”

Extrapolation validation (EV): a universal validation method for mitigating machine learning extrapolation risk

“…Machine learning (ML) has made impressive achievements in substance discovery, data analysis, and image processing over the past decades, accelerating advances in fields as numerous as earth & life science, 1–3 communications & transportation, 4–10 and chemistry & medicine. 11–20 As a spotlight to the field of chemistry, ML provides experimentalists with advice on selecting target molecules for synthesis by predicting physicochemical properties, 11–15 biological effects, 16–18 and reaction routes. 21–24 Although ML models are still not a complete substitute for expert intuition, 25 they are sufficiently sophisticated to recognize complex patterns beyond the reach of expert intuition to provide decision-making advice for major challenges in science and engineering, as multiple algorithms and different architectures for ML solutions emerge.…”