LATIDIA · Investigación

DeciCAD: Políticas de aprendizaje de acción basadas en el arnés para agentes de CAD

arXiv: 2610.21755v1Announce Type: new Resumen: Los agentes del modelo de lenguaje grande operan a través de políticas de acción de bucle cerrado cuyo espacio de decisión está definido por un arnés de agente. Investigamos cómo se configura este espacio operativo

La noticia

arXiv:2610.21755v1 Announce Type: new Abstract: Large Language Model agents operate through closed-loop action policies whose decision space is defined by an agent harness. We investigate how this operating space shapes agent behavior and policy learning in computer-aided design (CAD), a professional domain characterized by persistent state, geometric constraints, and long-horizon procedures. We instantiate this perspective in DeciCAD, a domain-native harness that exposes semantic CAD actions, decision-oriented observations, environment feedback, and persistent state transitions. Building on this harness, we develop a learning pipeline that extracts causal decision supervision from interaction trajectories and improves the policy through supervised fine-tuning (SFT) and geometry-grounded online reinforcement learning (RL). Across four CAD benchmarks, DeciCAD achieves