Id: arxiv/2611.11445 Url: https://arxiv.org/abs/2611.11445 Tags: quant-ph, cs.LG, physics.chem-ph Score: 0 Title: Alchemically Transferable Parameters for Variational Quantum Eigensolvers Domain: quant-ph Source: arxiv Authors: Jonas Motyl, Jakob S. Kottmann Reviews: 0 Subject: quant-ph Category: research Comments: 0 Disputed: false Division: physics Mentions: 0 Discussed: 0 Published: 2026-11-08 Page: https://delphi.science/r/arxiv/2611.11445 Machine-learned circuit parameters could remove one of the main bottlenecks of variational quantum eigensolvers: the classical optimization loop, with its typical barren plateaus, local minima, and costly energy and gradient evaluations, is replaced by a direct prediction. Prior work showed that machine-learning models paired with a structured circuit design can transfer across system size and structure within hydrogenic systems. The next step: Extending them to organic molecules would seem to require re-training on use-case-specific training sets. In this work, we show that this is not…