FutureWorlds: Learning Robotic World Models from Alternative Futures
cs.CV, cs.RO
Submitted: 2026-10-01
Updated: 2026-10-01
Comments: 32 pages, including references and appendix. Code: https://github.com/Alexander-wu/FutureWorlds
Code: https://github.com/Alexander-wu/FutureWorlds
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Thinking Fast and Slow with Deep Learning and Tree Search
- FitVid: Overfitting in Pixel-Level Video Prediction
- Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks
- Training Diffusion Models with Reinforcement Learning
- Visual Foresight: Model-Based Deep Reinforcement Learning for Vision-Based Robotic Control
- Unsupervised Learning for Physical Interaction through Video Prediction
- DreamDojo: A Generalist Robot World Model from Large-Scale Human Videos
- GigaWorld-1: A Roadmap to Build World Models for Robot Policy Evaluation
- Ctrl-World: A Controllable Generative World Model for Robot Manipulation
- Dream to Control: Learning Behaviors by Latent Imagination
- The Curious Case of Neural Text Degeneration
- LoRA: Low-Rank Adaptation of Large Language Models
- Vid2World: Crafting Video Diffusion Models to Interactive World Models
- WEAVER, Better, Faster, Longer: An Effective World Model for Robotic Manipulation
- Discovering Hidden Algebraic Structures via Transformers with Rank-Aware Beam GRPO
- DriftWorld: Fast World Modeling through Drifting
- Finite Scalar Quantization: VQ-VAE Made Simple
- World Simulation with Video Foundation Models for Physical AI
- Sequence Level Training with Recurrent Neural Networks
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models