LATIDIA · Robótica

Cómo capacitar mejor a los VLA: lecciones aprendidas del desafío REAL-I en ICRA 2026

arXiv:2609.13679v1 Announce Type: new Abstract: ¿Cómo pueden las políticas de robots aprender más eficazmente de un presupuesto de demostración fijo? El primer desafío de aprendizaje de IA incorporada en el mundo real (REAL-I) en ICRA 2026 examinó este que

La noticia

arXiv:2609.13679v1 Announce Type: new Abstract: How can robot policies learn more effectively from a fixed demonstration budget? The first Real-world Embodied AI Learning (REAL-I) Challenge at ICRA 2026 examined this question through simulation, real-robot evaluation, and an on-site final on a shared dual-arm humanoid platform. We describe the challenge tasks, data and deployment interfaces, and competition results, then compare the approaches contributed by NUS-CLEAR, RCL-Lab, and Deeptouch.ai. Their systems combined pretrained vision-language-action models and task-specific imitation policies with different strategies for data curation, staged adaptation, checkpoint selection, and action-space design. The team reports highlight the importance of adapting to the deployment environment while retaining prior capabilities, treating demonstration quality at