LATIDIA · Robótica

Manipulación bimanual de robots a través del aprendizaje en contexto de múltiples agentes

arXiv:2604.20348v3 Announce Type: replace Resumen: Los modelos de lenguaje grandes (LLM) han surgido como potentes motores de razonamiento para el control incorporado. En particular, el aprendizaje en contexto (ICL, por sus siglas en inglés) permite

La noticia

arXiv:2604.20348v3 Announce Type: replace Abstract: Large Language Models (LLMs) have emerged as powerful reasoning engines for embodied control. In particular, In-Context Learning (ICL) enables off-the-shelf, text-only LLMs to predict robot actions without any task-specific training while preserving their generalization capabilities. Applying ICL to bimanual manipulation remains challenging as the high-dimensional joint action space and tight inter-arm coordination constraints rapidly overwhelm standard context windows. To address this, we introduce BiCICLe (Bimanual Coordinated In-Context Learning), the first framework that enables standard LLMs to perform few-shot bimanual manipulation without fine-tuning. BiCICLe frames bimanual control as a multi-agent leader-follower problem, decoupling the action space into sequential, conditioned single-arm predictions. Evaluated on 13 tasks from