Auriga Technical Report: A Task-Agnostic Agent Harness for Zero-Shot Embodied Manipulation
cs.RO, cs.AI
Submitted: 2026-10-31
Updated: 2026-10-31
Project page: https://mmlab.hk/research/PhysicalRSI
Terminology
Sources
- SAM 3: Segment Anything with Concepts
- GaP: A Graph-as-Policy Multi-Agent Self-Learning Harness For Variational Automation Tasks
- RoboDojo: A Unified Sim-and-Real Benchmark for Comprehensive Evaluation of Generalist Robot Manipulation Policies
- In-Context Robot Learning with VLM Agents
- DynaHarness: A Dynamic Physical Harness for Self-Evolving Robot Agents
- The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement
- CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulation
- Systematically Exploring the Capabilities of GPT-6 Astra as Embodied Policies
- HexaShell: A Universal Harness for Robot Coding Agents in Simulation and the Real World
- Gemini Robotics 1.5: Pushing the Frontier of Generalist Robots with Advanced Embodied Reasoning, Thinking, and Motion Transfer
- Unified 4D World Action Modeling from Video Priors with Asynchronous Denoising
- Video Prediction Policy 2: Predict Better, Act Better
- Long-WAM: Scaling the Context of World-Action Models
- Agent as Policy for Robotic Manipulation
- RLDX-1 Technical Report
- Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning
- RynnBrain 1.1: Towards More Capable and Generalizable Embodied Foundation Model
- EmbodiedEvo: Self-Evolving Coding Agent For Physical Intelligence
- Code as Policies: Language Model Programs for Embodied Control
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