REVIEW 7 cited by
On learning higher-order cumulants in diffusion models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
To analyse how diffusion models learn correlations beyond Gaussian ones, we study the behaviour of higher-order cumulants, or connected n-point functions, under both the forward and backward process. We derive explicit expressions for the moment- and cumulant-generating functionals, in terms of the distribution of the initial data and properties of forward process. It is shown analytically that during the forward process higher-order cumulants are conserved in models without a drift, such as the variance-expanding scheme, and that therefore the endpoint of the forward process maintains nontrivial correlations. We demonstrate that since these correlations are encoded in the score function, higher-order cumulants are learnt in the backward process, also when starting from a normal prior. We confirm our analytical results in an exactly solvable toy model with nonzero cumulants and in scalar lattice field theory.
Forward citations
Cited by 7 Pith papers
-
Stochastic Quantization as Optimal Control
Stochastic quantization is re-expressed as finite-time optimal control, in which a learned Doob force plus exact path weights reach the Gibbs measure without waiting for equilibrium.
-
Diffusion Models for Sampling Near Criticality in Lattice Field Theories
Fully convolutional diffusion models trained on small lattices transfer to unseen larger volumes for 2D/3D phi^4 sampling across phases, matching or beating same-size training on most observables.
-
Exploring Generative Networks for Manifolds with Non-Trivial Topology
A GFlowNet-inspired diffusion sampler is proposed and shown, on toy and 2D lattice scalar models, to generate configurations across disconnected sectors that normalizing flows and plain diffusion models miss.
-
Diffusion models learn distributions generated by complex Langevin dynamics
Diffusion models reproduce the distributions sampled by complex Langevin dynamics in a Gaussian and a quartic toy model with complex mass.
-
Scaling of Stochastic Normalizing Flows in $\mathrm{SU}(3)$ lattice gauge theory
First demonstration that Stochastic Normalizing Flows inherit the linear-with-volume scaling of non-equilibrium MCMC in 4D SU(3) lattice gauge theory, with a factor-of-two efficiency gain.
-
Combining complex Langevin dynamics with score-based and energy-based diffusion models
Energy-based diffusion models trained on complex Langevin data produce an explicit energy function for the sampled distribution, enabling MCMC without re-simulation.
-
Diffusion models and stochastic quantisation in lattice field theory
Diffusion models, whose backward denoising step resembles stochastic quantisation, can learn from HMC data to generate configurations for 2D scalar lattice field theory.
Discussion (0). Continue with ORCID to comment.