# X-Foresight: A Joint Vision-Action Causal Forecasting Network via Predictive World Modeling

arXiv: https://arxiv.org/abs/2605.24892 · HTML: https://mlsanity.com/paper/2605.24892/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Embodied and robot world models, one of 11 directions in World models.
- Builds on 19 of its 22 references that are in the map.
- First posted on arXiv 24 May 2026.

## Placed in

- World models / [Embodied and robot world models](https://mlsanity.com/wm/embodied-and-robot-world-models/)
- Vision-language-action models / [World model-based VLA](https://mlsanity.com/vla/world-model-based-vla/)

## Builds on (most cited first, 19 of 19)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Towards Accurate Generative Models of Video: A New Metric & Challenges](https://mlsanity.com/paper/1812.01717/) (2018)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [WorldSimBench: Towards Video Generation Models as World Simulators](https://mlsanity.com/paper/2410.18072/) (2024)
- [Genie: Generative Interactive Environments](https://mlsanity.com/paper/2402.15391/) (2024)
- [Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion](https://mlsanity.com/paper/2506.08009/) (2025)
- [From Slow Bidirectional to Fast Autoregressive Video Diffusion Models](https://mlsanity.com/paper/2412.07772/) (2024)
- [Latent Diffusion Model without Variational Autoencoder](https://mlsanity.com/paper/2510.15301/) (2025)
- [Radial Attention: O(nlog n) Sparse Attention with Energy Decay for Long Video Generation](https://mlsanity.com/paper/2506.19852/) (2025)
- [Helios: Real Real-Time Long Video Generation Model](https://mlsanity.com/paper/2603.04379/) (2026)
- [One Layer Is Enough: Adapting Pretrained Visual Encoders for Image Generation](https://mlsanity.com/paper/2512.07829/) (2025)
- [X-World: Controllable Ego-Centric Multi-Camera World Models for Scalable End-to-End Driving](https://mlsanity.com/paper/2603.19979/) (2026)

## Built on by (most cited first, 1 of 1)

- [Situation Perception: A Necessary Primitive to Artificial Superintelligence](https://mlsanity.com/paper/2606.30481/) (2026)

## Abstract (arXiv)

Physical world knowledge resides mainly in videos. Equipping Vision-Language-Action (VLA) models with such knowledge is fundamental for safe and generalizable planning. Predictive world modeling enables VLA to internalize physical dynamics and long-term causality by predicting future video from past observations. However, naive next-frame prediction faces two challenges: 1) unlike semantically distinct text tokens, video tokens are low-entropy and redundant, causing prediction to degenerate into trivial extrapolation. 2) world modeling poses a temporal dilemma: dense prediction captures instantaneous dynamics, but cannot efficiently model long-horizon causality. To learn world knowledge effectively, we introduce X-Foresight, a predictive world model integrated directly into the VLA architecture to jointly learn world modeling and real-time action control. At its core lies a long-horizon chunk-wise auto-regressive strategy that addresses both challenges: by predicting semantically distant chunks rather than adjacent frames, it escapes trivial extrapolation, while preserving dense intra-chunk frames for instantaneous dynamics and sparse inter-chunk transitions for long-term causality. A curriculum learning schedule progressively extends prediction horizons and stabilizes long-horizon training. To capture long-term causality effectively, we present temporal importance sampling, which concentrates supervision on safety-critical chunks identified by ego-motion and behavioral signals. We further delegate photorealistic synthesis to a diffusion-based multi-view renderer, improving photorealistic appearance. Comprehensive experiments demonstrate that X-Foresight significantly outperforms VLA baselines in planning performance while maintaining strong generative fidelity, establishing a robust paradigm for world-knowledge-driven autonomous systems.