DROID

The 'DROID' dataset/benchmark is used to evaluate visual prediction models for embodied control by providing a collection of robotic manipulation tasks and corresponding visual states.

Papers using DROID (10)

V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning2025 · 4 citesSWEET: Sparse World Modeling with Image Editing for Embodied Task Execution2026CAPE: Contrastive Action-conditioned Parallel Encoding for Embodied Planning2026ROSER: Few-Shot Robotic Sequence Retrieval for Scalable Robot Learning2026Sparse Autoencoders Reveal Interpretable and Steerable Features in VLA Models2026Persistent Robot World Models: Stabilizing Multi-Step Rollouts via Reinforcement Learning2026RAAP: Retrieval-Augmented Affordance Prediction with Cross-Image Action Alignment2026BridgeV2W: Bridging Video Generation Models to Embodied World Models via Embodiment Masks2026Ctrl-World: A Controllable Generative World Model for Robot Manipulation2025EmbodiedMAE: A Unified 3D Multi-Modal Representation for Robot Manipulation2025