New AI System Teaches Itself to Control Robots Without Human Help
This could lead to cheaper, smarter robots that learn on their own.
EliteTable is a self-improving program-search algorithm guided by large language models (LLMs) that uses a fixed-capacity table of high-scoring solutions to guide new proposals, edits of existing solutions, and remixes of several solutions. Performance was evaluated across six Gymnasium control tasks, each with ten generations and fifty candidate slots per generation. Each task was evaluated across ten independently initialized searches.
Candidate-score medians rose between the first and final generations in all six tasks, and selected-winner search-score medians rose in five. In CartPole-v1, the selected winner reached the search-score ceiling in the first generation of all ten searches. All ten paired held-out endpoints improved in four of the six tasks.
BipedalWalker-v3 and CarRacing-v3 improved without reaching their reporting targets. Six of ten Acrobot-v1 searches declined on held-out episodes, and the median paired change was negative. The results document improvement during program search and the transfer of selected-winner gains to new episodes in some tasks. Comparisons and ablations are needed to establish EliteTable's relative performance and the contributions of its components.
- EliteTable is an AI that writes and improves robot control programs by itself, without human intervention.
- It could drastically reduce the time and cost of programming robots for tasks like assembly or delivery.
- The system still needs powerful computers and works best in simulation, so real-world use is a few years away.
Why It Matters
This could lead to cheaper, more adaptable robots, impacting jobs in manufacturing, logistics, and service industries.