MESA: An Evaluation Framework for Compositional, Semantic, and Spatial Generalization in Robotics
cs.RO, cs.AI
Submitted: 2026-11-03
Updated: 2026-11-03
Code: https://github.com/oogle-deepmind/mujoco_menagerie
Project page: https://pairlab.github.io/MESA
Terminology
Sources
- RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control
- OpenVLA: An Open-Source Vision-Language-Action Model
- FAST: Efficient Action Tokenization for Vision-Language-Action Models
- Gemini Robotics: Bringing AI into the Physical World
- GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
- MolmoAct: Action Reasoning Models that can Reason in Space
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- Gemini Robotics 1.5: Pushing the Frontier of Generalist Robots with Advanced Embodied Reasoning, Thinking, and Motion Transfer
- GR-3 Technical Report
- SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics
- ${\pi}_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities
- World Action Models are Zero-shot Policies
- mimic-video: Video-Action Models for Generalizable Robot Control Beyond VLAs
- MolmoAct2: Action Reasoning Models for Real-world Deployment
- Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning
- RoboCasa: Large-Scale Simulation of Everyday Tasks for Generalist Robots
- BEHAVIOR-1K: A Human-Centered, Embodied AI Benchmark with 1,000 Everyday Activities and Realistic Simulation
- PolaRiS: Scalable Real-to-Sim Evaluations for Generalist Robot Policies
- RobotArena $\infty$: Scalable Robot Benchmarking via Real-to-Sim Translation
- REALM: A Real-to-Sim Validated Benchmark for Generalization in Robotic Manipulation
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation