AI RESEARCH
Kintsugi: Learning Policies by Repairing Executable Knowledge Bases
arXiv CS.LG
•
ArXi:2605.09487v1 Announce Type: new Modern embodied agents achieve impressive performance, but their task knowledge is often d in neural weights, latent state, or prompt-bound memory, making individual policy knowledge difficult to inspect, validate, recombine, and reuse. We