Search papers, labs, and topics across Lattice.
The paper introduces Moment Guided Diffusion (MGD), a novel approach to sample maximum entropy distributions by guiding moments toward prescribed values using a stochastic differential equation within the stochastic interpolant framework. This method addresses the limitations of traditional maximum entropy methods (slow mixing) and diffusion models (limited theoretical guarantees) when generating samples from limited information. The authors demonstrate that MGD converges to the maximum entropy distribution in the large-volatility limit and provide a tractable entropy estimator, showcasing its effectiveness on financial time series, turbulent flows, and cosmological fields using wavelet scattering moments.
Escape the MCMC slowdown: this diffusion-based approach efficiently samples maximum entropy distributions by directly guiding moments, offering a practical alternative for high-dimensional uncertainty quantification.
Generating samples from limited information is a fundamental problem across scientific domains. Classical maximum entropy methods provide principled uncertainty quantification from moment constraints but require sampling via MCMC or Langevin dynamics, which typically exhibit exponential slowdown in high dimensions. In contrast, generative models based on diffusion and flow matching efficiently transport noise to data but offer limited theoretical guarantees and can overfit when data is scarce. We introduce Moment Guided Diffusion (MGD), which combines elements of both approaches. Building on the stochastic interpolant framework, MGD samples maximum entropy distributions by solving a stochastic differential equation that guides moments toward prescribed values in finite time, thereby avoiding slow mixing in equilibrium-based methods. We formally obtain, in the large-volatility limit, convergence of MGD to the maximum entropy distribution and derive a tractable estimator of the resulting entropy computed directly from the dynamics. Applications to financial time series, turbulent flows, and cosmological fields using wavelet scattering moments yield estimates of negentropy for high-dimensional multiscale processes.
