AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

Learning Hamiltonian Flow Maps: Mean Flow Consistency for Large-Timestep Molecular Dynamics

Winfried Ripken, Michael Plainer, Gregor Lied, Thorben Frank, Oliver T. Unke, Stefan Chmiela, Frank Noe, Klaus Robert Muller

ICML 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Stochastic population dynamics · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase HFM integration timestep from 0.5 fs to 9 fs (ablation)HFM · Aspirin NVTh(r) MAE: 0.026 -> 0.046Hurt
increase HFM integration timestep from 0.5 fs to 9 fs (ablation)HFM · Ethanol NVTh(r) MAE: 0.075 -> 0.115Hurt
replace MLFF+Velocity Verlet with HFM at 0.5 fsMLFF · Aspirin NVTh(r) MAE: 0.030 -> 0.026Improved
use HFM instead of Velocity Verlet for N-body rollout at coarse discretizationGravitational N-body (100 bodies) · N-body t=3 to t*=4MSE: 4x fewer steps than VV for same accuracyImproved
use HFM with dt=9 fs instead of MLFF+VV at 0.5 fs for fixed step budgetParacetamol NVE · Paracetamol NVE explorationstate space explorationImproved
distill CG ScoreMD into HFM with dt=45 fsChignolin CG · Chignolin NVTspeedup vs reference force field: exceeding 20xImproved
use Nose-Hoover thermostat instead of LangevinHFM · NVT simulationssimulation artifactsHurt
train on sparse 256 samples instead of 85k samplesHFM · Paracetamol NVTfree energy surface and spectrumNo clear effect
raise alanine dipeptide dt to about 10 fsHFM · Alanine dipeptide NVTrollout stabilityHurt
sample dt=0 in 75% of training casesHFM training · MLFF datasetstraining stabilityImproved
use stop-gradient on consistency targetHFM training · training costtraining overhead: approximately 30% overhead vs MLFFImproved
train HFM on biased data with a missing conformational modeHFM · MD22 Ac-Ala3-NHMe NVTfree energy landscapeImproved

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