# Flow Matching in Feature Space for Stochastic World Modeling

Preprint · Francois Porcher, Nicolas Carion, Karteek Alahari, Shizhe Chen

- Archive id: arxiv:2606.29059
- Published: 2026-06-27
- arXiv: https://arxiv.org/abs/2606.29059
- Citations: 0
- Source URL: https://arxiv.org/abs/2606.29059
- PDF: https://arxiv.org/pdf/2606.29059
- Provenance: arxiv

## Abstract

World modeling requires forecasting uncertain futures while preserving information useful for downstream perception. Existing visual world models often struggle to satisfy both goals: VAE-based stochastic models operate in low-dimensional reconstruction latents, which can limit perception performance, while deterministic predictors using strong pretrained features collapse multimodal futures into a single blurry mean. In this work, we propose FlowWM, a stochastic world model that performs flow matching directly within pretrained feature space (e.g., DINOv3). This is challenging because pretrained features are substantially high-dimensional, making standard diffusion recipes suboptimal. To address this, we investigate the design choices needed for feature-space flow matching and introduce a differentiable one-step projection mechanism that enables efficient training with temporal consistency and task-driven objectives. We evaluate FlowWM on two benchmarks: a synthetic benchmark for systematic evaluation of accuracy and diversity, and a real-world benchmark FuturePerception. FlowWM improves perception performance, mode coverage, and horizon robustness, validating our proposed design for stochastic world modeling in high-dimensional feature spaces.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [Generative & Multimodal Models](https://space-kitty.com/topics/generative-multimodal)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.