# Simulation Distillation: Pretraining World Models in Simulation for Rapid Real-World Adaptation

Source: [Simulation Distillation: Pretraining World Models in Simulation for Rapid Real-World Adaptation](<https://www.sota2.com/research/paper/69ba31710636e5bf4afc065c>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-05-14T05:00:13.174Z

Authors: Jacob Levy, Tyler Westenbroek, Kevin Huang, Fernando Palafox, Patrick Yin, Shayegan Omidshafiei, Dong-Ki Kim, Abhishek Gupta, David Fridovich-Keil

Published: 2026-03-16T18:00:23.000Z

## Abstract

Robot learning requires adaptation methods that improve reliably from limited, mixed-quality interaction data. This is especially challenging in long-horizon, contact-rich tasks, where end-to-end policy finetuning remains inefficient and brittle. World models offer a compelling alternative: by predicting the outcomes of candidate action sequences, they enable online planning through counterfactual reasoning. However, training action-conditioned robotic world models directly in the real world requires diverse data at impractical scale. We introduce Simulation Distillation (SimDist), a framework that uses physics simulators as a scalable source of action-conditioned robot experience. During pretraining, SimDist distills structural priors from the simulator into a world model that enables planning from raw real-world observations. During real-world adaptation, SimDist transfers the encoder, reward model, and value function learned in simulation, and updates only the latent dynamics model using real-world prediction losses. This reduces adaptation to supervised system identification while preserving dense, long-horizon planning signals for online improvement. Across contact-rich manipulation and quadruped locomotion tasks, SimDist rapidly improves with experience, while prior adaptation methods struggle to make progress or degrade during online finetuning. Project website and code: https://sim-dist.github.io

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Quadruped Locomotion on Slippery Slope real-world evaluation](<https://www.sota2.com/research/sota/quadruped-locomotion-on-slippery-slope-real-world-evaluation.md>): Quadruped Locomotion; Slippery Slope real-world evaluation; Forward Progression: 1.82.

- [Quadruped Locomotion on Foam real-world (evaluation)](<https://www.sota2.com/research/sota/quadruped-locomotion-on-foam-real-world-evaluation.md>): Quadruped Locomotion; Foam real-world (evaluation); Forward Progression Distance: 3.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.15759>)

[PDF](<https://arxiv.org/pdf/2603.15759.pdf>)
