{"id":"arxiv/2607.22211","url":"https://arxiv.org/abs/2607.22211","tags":["math.NA","cs.NE"],"score":0,"title":"Closed-Loop Generative Selection: Convergence, Memory, and Noisy Oracles","domain":"math.NA","source":"arxiv","authors":["Konstantin Fackeldey","Christof Schütte"],"reviews":0,"subject":"math.NA","summary":"Closed-loop generative selection has become a workhorse of computational drug discovery: a learned generative model proposes candidate molecules, a fitness oracle scores them, the best are kept, and the model is retrained on this elite set before the next round. Despite its wide use, the method has lacked a rigorous convergence theory, largely because retraining the model each round breaks the Markov property on which classical evolutionary-algorithm analysis relies. We develop a self-contained theory of convergence and expected running time for this class of algorithms. By recovering a…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-24","reactions":{}}