Felix Heide

How should a world model plan when the goal is out of sight? 🗺️🤔

Latent world models, like JEPA, commit to a single future — being deterministic by design! At #NeurIPS2026, we present EpicWorldModel 🌍✨: a stochastic latent world model that samples many futures and uses their disagreement as an exploration bonus in planning.

It explores where it’s uncertain 🔍 and commits once the goal comes into view🎯.

We achieve +22% over LeWorldModel on Visual PointMaze Giant 🏆, and in robotic navigation with a mostly hidden goal, success goes from 17 to 83% 🤖📈

Great work led by Bowen Feng, Julian Ost 🙌

Project Page: epicworldmodel.com
Paper: arxiv.org/abs/2610.05996

Can robots learn to reason and act with NO human-annotated language or reasoning labels?

We find that VLMs can teach themselves to understand goal and perform robot actions! We call this Action Self-Distillation (ASD) for language-goal-conditioned reasoning and navigation. ASD outperforms all VLAs trained on hand-annotated data that we tested! 🤖 #CoRL2026

Project and Paper: asd-nav.github.io

Super fun project lead by @BowenFengbf

Real-world scenes are highly repetitive 🧱🌿: bricks, blades of grass, leaves, and many other structures appear again and again. At #NeurIPS2026, we present SCION, which learns a compact vocabulary of reusable 3D primitives and instantiates them throughout a scene, capturing its underlying compositional structure. This brings learned photorealistic 3D Scenes down to the ~1 MB range 🤏 while introducing natural structures for editing and animation 🎬.

Great work by @williamkoch314, @amogh7joshi, @cvachha, and Cheng Zheng.

You'll see the authors in Sydney this December!

Project Page: light.princeton.edu/SCION
Paper: arxiv.org/abs/2610.02322

Super excited to share this unedited fully autonomous truck ride! We're seeing strong zero-shot planning capabilities emerge from self- and mixed-play RL training at scale in complex situations. Our stack has not been trained on scenarios from this unseen route, and the generalization capabilities are tremendous!

Proud of the @torc_robotics team for their work on this frontier architecture for autonomous driving.

ScenarioControl (ECCV 2026) enables end-to-end controllable generative simulation 🚗✨for autonomous driving, allowing users to generate diverse traffic scenarios with fine-grained actor control while preserving realistic scene dynamics.

We have released the code, and we’re looking forward to seeing what the community builds with it!

🔗 Check out our project page for the paper & code:
…ceton-computational-imaging.github.io/ScenarioContro…

👏 Thanks to the first authors @Lili_Gao_ , @Yanbo_Xu_ , @WilliamKoc61157