Convergence Mechanisms of Generative Models in Molecular Conformational Sampling
The convergence mechanism is established as a practical design principle for molecular generative sampling, clarifying when stochastic diffusion provides robustness and when deterministic transport requires higher representational capacity.