AI Paper Atlas

Research map / Optimization / Optimizers

Macrocanonical Generator Networks: data-efficient neural surrogates for amortized physics simulation

Niall Jeffrey, Benjamin Wandelt

NeurIPS 2026 · Optimization › Optimizers · approach: Natural gradient and Fisher approximations · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces macrocanonical generator networks, which learn fast, data-efficient neural surrogates for amortized simulation of stationary physical processes by matching multiscale statistics of the target process.

Problem

Scientific simulations are computationally expensive, which limits their use in simulation-based inference, uncertainty quantification, and decision making.

Approach

Learns neural surrogates for amortized simulation of stationary physical processes by matching multiscale statistics of a target process, inspired by microcanonical maximum entropy (generator-network) ideas.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
pretrain SBI encoder on macrocanonical-generated synthetic dataResNet encoder + Gaussian neural density estimator · Quijote Omega_DE inferencevalidation NLL: -0.93 -> -1.69Improved
train from scratch on real simulations instead of synthetic pretrainingSBI encoder · Quijote Omega_DE inferencevalidation NLLHurt
remove the residual shortcut (ablation)macrocanonical generator · MHD and cosmic-websample quality / collapseHurt
use large learning rate eta=1e-2 instead of small (ablation)macrocanonical generator · MHD and cosmic-websample quality / collapseHurt
replace circular padding with zero padding (ablation)macrocanonical generator · MHD and cosmic-webstationarityHurt
train on more fields (1, 10, 100) (ablation)macrocanonical generator · MHD turbulencestatistics vs held-outImproved
use macrocanonical training instead of WGAN-GP, flow matching or DDPMgenerator · single training fieldmemorizationImproved
match features in expectation instead of per-sample (microcanonical)generator · Quijote slicesper-realization fluctuationsImproved

Experimental setup

Builds on: microcanonical maximum entropy

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy