Ten Advances in Theoretical and Mathematical Chemistry
Year: 2026
Type: article
Abstract: This work presents ten AI-generated or substantially AI-assisted studies at the interfaceof theoretical chemistry, mathematical physics, spectral theory, many-body quantummechanics, electronic-structure theory, statistical mechanics, molecular dynamics, andrigorous computation. Human-authored editorial material provides the scientific framing,provenance information, and methodological oversight.The objective is not simply to reproduce conventional computational-chemistry results,nor to claim that artificial intelligence has autonomously solved ten major problems. Rather,the volume tests how far contemporary AI can reformulate central questions of theoreticalchemistry as mathematically controlled problems with explicit assumptions, identifiablelogical dependencies, and conclusions that can in principle be independently audited.The chapters address stability of Coulomb matter and molecular ground states, theBorn–Oppenheimer approximation, chemical bonding, density functional theory, electroniccorrelation, reaction dynamics and resonances, the quantum-to-classical limit, equilibriumthermodynamics, chemical kinetics, and molecular transport. Their scientific status is deliberatelyheterogeneous: some chapters reconstruct and organize established mathematicalresults, others develop controlled approximation frameworks or rigorous inequalities, andothers formulate explicit certification targets or identify conditions under which strongerconclusions would follow.A recurring distinction throughout the volume is between numerical evidence and mathematicalcertification. Whenever possible, scientific claims are reduced to explicit inequalities,spectral statements, variational principles, existence results, asymptotic estimates, limitingtheorems, or computable certificates.The broader methodological question is whether AI systems can contribute not only tocalculation and exposition, but also to the construction, testing, and adversarial auditingof mathematically controlled scientific arguments. AI-generated reasoning is not treated asindependent verification; the scientific status of the results ultimately depends on humanspecialist review and, where applicable, formal or certified computational verification.
Authors: Sienicki, Krzysztof, Sienicki, Mikołaj
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7208822823