NVIDIA AI Introduces ASPIRE: A Self-Improving Robotics Framework Reaching 31% Zero-Shot on LIBERO-Pro Long Tasks

NVIDIA researchers have introduced ASPIRE, a framework that enables robots to autonomously write and refine control programs before storing successful solutions in a reusable skill library. By iteratively self-improving through trial and error, the system demonstrated improved performance on complex, long-horizon tasks in the LIBERO-Pro benchmark. This approach aims to reduce the need for extensive manual programming by allowing robots to build their own functional knowledge base for executing unseen tasks.