Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Arcadia: Toward a Full-Lifecycle Framework for Embodied Lifelong Learning
Minghe Gao, Juncheng Li, Yuze Lin, Xuqi Liu, Jiaming Ji, Xiaoran Pan, Zihan Xu, Xian Li, Mingjie Li, Wei Ji, Rong Wei, Rui Tang and 6 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add real-world feedback loop (ablation) | Arcadia · VLN-CE-Isaac | SR: 48.7 -> 50.1 | Improved |
| add real-world feedback loop (ablation) | Arcadia · BridgeData V2 | success rate: 47.3 -> 52.4 | Improved |
| use dynamic generative scene reconstruction instead of retrieval-based (ablation) | Arcadia · LIBERO | success rate: 81.4 -> 87.2 | Improved |
| use task-aligned autonomously collected data instead of static ScaleVLN/RLBench data (ablation) | Arcadia · LIBERO | success rate: 72.9 -> 87.2 | Improved |
| use task-aligned autonomously collected data instead of static data (ablation) | Arcadia · VLN-CE-Isaac | SR: 43.0 -> 50.1 | Improved |
| use dense structured feedback instead of binary success/failure (ablation) | Arcadia · VLN-CE-Isaac | SR: 48.8 -> 50.1 | Improved |
| add joint VLN and VLA training on shared backbone (ablation) | Arcadia · VLN-CE-Isaac | SR: 49.8 -> 50.1 | No clear effect |
| add joint VLN and VLA training on shared backbone (ablation) | Arcadia · LIBERO | success rate: 87.0 -> 87.2 | No clear effect |
| replace NaVILA external data with Arcadia self-collected data (no feedback) | NaVILA · R2R Val-Unseen | SR: 51.6 -> 54.2 | Improved |
| replace OpenVLA training data with Arcadia simulation data (no feedback) | OpenVLA · LIBERO-10 | success rate: 53.7 -> 74.0 | Improved |
| apply Arcadia in real-world zero-shot navigation | NaVILA · Real-world navigation | success rate: 13% -> 46% | Improved |
Closest papers by meaning
- SimWorld Studio: Automatic Environment Generation with Evolving Coding Agent for Embodied Agent Learning (NeurIPS 2026 · similarity 0.88)
- Universal Actions for Enhanced Embodied Foundation Models (CVPR 2025 · similarity 0.87)
- Embodied-R1.5: Evolving Physical Intelligence via Embodied Foundation Models (NeurIPS 2026 · similarity 0.87)
- SEEA-R1: Tree-Structured Reinforcement Fine-Tuning for Self-Evolving Embodied Agents (NeurIPS 2025 · similarity 0.87)
- Self-Improving Embodied Foundation Models (NeurIPS 2025 · similarity 0.87)
- Learning 3D Persistent Embodied World Models (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
