Gradient-Free Sampling from Generative Models via Stochastic Bounded Extremum Seeking

We introduce a sampling approach for energy- and score-based generative models that requires no gradient evaluations of the model. Replacing the drift term that would normally contain the score nablax log pθ(bfx) with a high-frequency dithered cosine of the model's textitvalue, √αω,cos(ω t + k log pθ(bfx)), produces, in the high-frequency averaging limit, Langevin Markov chain Monte Carlo for energy-based models and the reverse-time SDE of score-based diffusion. We prove that trajectories of the dithered Ito SDE converge to those of the target SDE, driven by the same Brownian motion, uniformly on compact time intervals in probability, by an averaging argument that extends bounded extremum seeking (ES) to Ito processes, with an explicit O(ω-1/2) mean-square rate under global bounds. The approach is not confined to smooth targets: it extends to C1,1 energies with discontinuous curvature (without ellipticity requirement) and to Sobolev energies whose Hessians exist only off measure zero sets; for Lipschitz energies with gradient kinks the averaged limit remains well posed; the Krylov-R"ockner integrability class is the boundary of provability. The approach provides a hard textita priori bound on the per-step update rate and applies to explicitly time-varying targets on finite horizons. Gradient-free pixel-space sampling is not competitive with well-tuned backpropagation-based samplers at practical evaluation budgets; the regime where the approach offers an advantage is latent-space sampling when the model is a black box and the target drifts in time. We demonstrate latent-space tracking for time-varying images on CelebA-HQ (256×256) from limited 1D projection measurements and latent-space EBM sampling on CIFAR-10.