Stein Variational Gradient Descent
PulseAugur coverage of Stein Variational Gradient Descent — every cluster mentioning Stein Variational Gradient Descent across labs, papers, and developer communities, ranked by signal.
-
New Riesz-Kernel SVGD method tackles infinite self-interaction · 2 sources tracked
Researchers have developed a new method for Stein Variational Gradient Descent (SVGD) that addresses the issue of infinite self-interaction when using singular Riesz kernels. This modified approach, termed periodic Ries…
-
New rates improve Stein Variational Gradient Descent convergence
Researchers have developed new finite-particle convergence rates for the Stein Variational Gradient Descent (SVGD) algorithm. These advancements provide improved theoretical understanding for SVGD's performance in Kerne…
-
New paper explores uniform-in-time propagation-of-chaos for SVGD
Researchers have published a paper detailing uniform-in-time propagation-of-chaos for Stein Variational Gradient Descent (SVGD). The study introduces a cutoff strategy for broad distributional metrics, yielding propagat…
-
New research advances flow matching models with theoretical and algorithmic improvements
Researchers have developed new theoretical foundations and practical algorithms for flow matching models, a type of generative model. One paper establishes convergence guarantees for neural network-parameterized conditi…
-
New research advances generative models for efficiency and evaluation
Several recent research papers explore advancements in generative models, focusing on improving their efficiency, evaluability, and alignment. One paper proposes a new framework for weighted sampling using score-based g…
-
SVGD mean-field convergence rates quantified in new paper
Researchers have established quantitative local convergence rates for the mean-field limit of Stein Variational Gradient Descent (SVGD). This deterministic particle method is used for sampling from probability measures …