Markov spin models for image generation : explicit large deviations with respect to the number of pixels

Abstract

For the discrete-time or the continuous-time Markov spin models for image generation when each pixel $n=1,..,N$ can take only two values $S_n=\pm 1$, the finite-time forward propagator depends on the initial and on the final configurations of the $N$ spins only via a single global variable, namely the extensive overlap that counts the number of spins that have the same value or not in the two configurations. The joint probability distribution of the overlap and of the magnetization during the forward noising dynamics can be written for any finite number $N$ of pixels and in the limit $N\to +\infty$ to extract the large deviations properties. The consequences for the backward reconstructive dynamics are then analyzed for various initial conditions, namely (i) a single image (ii) a mixture of two images (iii) when the initial condition corresponds to the Curie-Weiss mean-field ferromagnetic model in the microcanonical ensemble, as a simple analog of the manifold-hypothesis concerning continuous generative diffusion models.

v2= revised version with new discussions and some new material (30 pages)

References in corpus (60)

• The large deviation approach to statistical mechanics
• Non equilibrium steady states: fluctuations and large deviations of the density and of the current
• Macroscopic fluctuation theory
• Dynamic first-order phase transition in kinetically constrained models of glasses
• Fluctuation theorems for stochastic dynamics
• First-order dynamical phase transition in models of glasses: an approach based on ensembles of histories
• Thermodynamic formalism for systems with Markov dynamics
• Direct evaluation of large-deviation functions
• Nonequilibrium Markov processes conditioned on large deviations
• Fluctuation Theorems
• Large deviations and ensembles of trajectories in stochastic models
• Nonequilibrium microcanonical and canonical ensembles and their equivalence
• Simulating rare events in dynamical processes
• Introduction to dynamical large deviations of Markov processes
• Unravelling the large deviation statistics of Markovian open quantum systems
• Aspects of non-equilibrium in classical and quantum systems: slow relaxation and glasses, dynamical large deviations, quantum non-ergodicity, and open quantum dynamics
• Ergodicity and large deviations in physical systems with stochastic dynamics
• Effective interactions and large deviations in stochastic processes
• Chaotic properties of systems with Markov dynamics
• Construction of a Coordinate Bethe Ansatz for the asymmetric simple exclusion process with open boundaries
• The Physicist's Companion to Current Fluctuations: One-Dimensional Bulk-Driven Lattice Gases
• Current large deviations for driven periodic diffusions
• Making rare events typical in Markovian open quantum systems
• A minimal model of dynamical phase transition
• Transition probabilities and dynamic structure factor in the ASEP conditioned on strong flux
• Large deviations and dynamical phase transitions in stochastic chemical networks
• Dynamical phase transition in drifted Brownian motion
• Large deviation approach to nonequilibrium systems
• Large deviations conditioned on large deviations I: Markov chain and Langevin equation
• Density profiles, dynamics, and condensation in the ZRP conditioned on an atypical current
• Microscopic structure of shocks and antishocks in the ASEP conditioned on low current
• Generic Dynamical Phase Transition in One-Dimensional Bulk-Driven Lattice Gases with Exclusion
• Second-order dynamic transition in a p=2 spin-glass model
• Effective driven dynamics for one-dimensional conditioned Langevin processes in the weak-noise limit
• Diffusions conditioned on occupation measures
• Large deviations conditioned on large deviations II: Fluctuating hydrodynamics
• Large-deviation theory for a Brownian particle on a ring: a WKB approach
• Large deviations at level 2.5 for Markovian open quantum systems: quantum jumps and quantum state diffusion
• Spectral theory of fluctuations in time-average statistical mechanics of reversible and driven systems
• Dynamical large deviations of reflected diffusions
• Bethe Ansatz for the Weakly Asymmetric Simple Exclusion Process and phase transition in the current distribution
• Thermodynamic formalism and large deviation functions in continuous time Markov dynamics
• Exact Distributions of Currents and Frenesy for Markov Bridges
• Generative diffusion in very large dimensions
• Microscopic Fluctuation Theory (mFT) for interacting Poisson processes
• Dynamical large deviations of linear diffusions
• Large Deviations for the density and the current in Non-Equilibrium-Steady-States on disordered rings
• Large deviations of the Lyapunov exponent in 2D matrix Langevin dynamics with applications to one-dimensional Anderson Localization models
• Large deviations of currents in diffusions with reflective boundaries
• Large Deviations for dynamical fluctuations of Open Markov processes, with application to random cascades on trees
• Nonequilibrium diffusion processes via non-Hermitian electromagnetic quantum mechanics with application to the statistics of entropy production in the Brownian gyrator
• Revisiting the Ruelle thermodynamic formalism for Markov trajectories with application to the glassy phase of random trap models
• Conditioned stochastic particle systems and integrable quantum spin systems
• Statistical physics of long dynamical trajectories for a system in contact with several thermal reservoirs
• Effective Hamiltonians and Lagrangians for conditioned Markov processes at large volume
• Anomalous dynamical large deviations of local empirical densities and activities in the pure and in the random kinetically-constrained East Model
• Conditioned diffusion processes with an absorbing boundary condition for finite or infinite horizon
• Dynamical large deviations of diffusions
• Strongly constrained stochastic processes: the multi-ends Brownian bridge
• Large deviations for the Pearson family of ergodic diffusion processes involving a quadratic diffusion coefficient and a linear force

Cited by in corpus (1)

• Convergence properties of Markov models for image generation with applications to spin-flip dynamics and to diffusion processes