LATIDIA · Robótica

Unificación de la invención de predicados profundos con modelos de cimentación preentrenados

arXiv: 2512.17992v2Announce Type: replace Resumen: Las tareas robóticas de horizonte largo son difíciles debido a los espacios continuos de acción de estado y la escasa retroalimentación. Los modelos de mundo simbólico ayudan a descomponer las tareas en predicados discretos

La noticia

arXiv:2512.17992v2 Announce Type: replace Abstract: Long-horizon robotic tasks are hard due to continuous state-action spaces and sparse feedback. Symbolic world models help by decomposing tasks into discrete predicates that capture object properties and relations. Existing methods learn predicates either top-down, by prompting foundation models without data grounding, or bottom-up, from demonstrations without high-level priors. We introduce UniPred, a bilevel learning framework that unifies both. UniPred uses large language models (LLMs) to propose predicate effect distributions that supervise neural predicate learning from low-level data, while learned feedback iteratively refines the LLM hypotheses. Leveraging strong visual foundation model features, UniPred learns robust predicate classifiers in cluttered scenes. We further propose a predicate