.png){kind=icon}
.png){kind=icon}
Return
SCORE-BASED DETERMINISTIC DENSITY SAMPLING
.png){kind=icon}
PRE
AI
V
P
H
DOI：10.3934/cpaa.2026028.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We propose a deterministic sampling framework using Score-Based Transport Modeling for sampling an unnormalized target density pi given only its score del log pi. Our method approximates the Wasserstein gradient flow on KL(f(t)parallel to pi) by learning the time-varying score del log f(t) on the fly using score matching. While having the same marginal distribution as Langevin dynamics, our method produces smooth deterministic trajectories, resulting in monotone noise-free convergence. We prove that our method dissipates relative entropy at the same rate as the exact gradient flow, provided sufficient training. Numerical experiments validate our theoretical findings: our method converges at the optimal rate, has smooth trajectories, and is often more sample efficient than its stochastic counterpart. Experiments on high-dimensional image data show that our method produces high-quality generations in as few as 15 steps and exhibits natural exploratory behavior. The memory and runtime scale linearly in the sample size.
Keywords:
Deterministic sampling
score-based transport modeling
Wasserstein gradient flow
relative entropy
Fisher information
log-Sob olev inequality
annealing
neural network
neural tangent kernel
Journal
C
IF:
0.9
Papers:
88
Citations:
0
Organization
Cited Papers