DashVMC: Real-Time Discrete World Model Control in Geometry Dash
On the map: World models → Latent dynamics models and model-based RL also fits: Latent predictive world models
Abstract (arXiv)
World-model agents are usually evaluated in simulators that can wait for the policy; live games impose the opposite constraint, requiring capture, prediction, and action before the next frame. We present DashVMC, which learns a compact, action-conditioned world model from approximately two hours of recorded Geometry Dash gameplay. To test whether the learned dynamics are actionable, a controller is initialized by behavioural cloning (BC) and refined with Proximal Policy Optimization (PPO) entirely in frozen-model rollouts, without further interaction with the live game. Across three controller seeds, the refined policies survive longer than their BC initializations on all three official levels and a held-out community layout. At deployment, the baseline skips visual generation and sustains a 60-Hz capture-to-action loop on a consumer GPU. Action-conditioned continuations and rollout diagnostics show that the model remains useful for control despite imperfect long-horizon fidelity.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Latent dynamics models and model-based RL, one of 11 directions in World models. Builds on 6 of its 17 references that are in the map. First posted on arXiv 30 Sep 2026.
Builds on
6 of 17 references.
References · 6
- 2017Proximal Policy Optimization Algorithmscited by 14,441
- 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
- 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 1,916
- 2022Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videoscited by 338
- 2026iFSQ: Improving FSQ for Image Generation with 1 Line of Codecited by 4
- 2025Cloning Deterministic Worlds: The Critical Role of Latent Geometry in Long-Horizon World Modelscited by 3
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Valerant: An Automatic Navigable Game Map Generator via Action-Conditioned World Model Exploration
- 2026GameWAM: A World Action Model for Video Games
- 2026WorldSimProbe: Diagnosing Simulator Faithfulness in Action-Conditioned World Models for Embodied Manipulation
- 2026PlayWorld: Benchmarking World Models with Agent Players over Long-Horizon Objectives
- 2026Matrix-Game 3.5: Enhancing Real-Time Streaming Interactive World Models with Patch Memory
- 2020Mastering Atari with Discrete World Models
- 2026Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models
- 2026WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models