"Instead of improving an agent against a fixed test, we let the evaluation evolve alongside the agent," Alex adds. "As the agent gets better, the evaluation also gets harder, and the bar keeps rising."
cst.cam.ac.uk
The Red Queen hypothesis - a new way forward for self-improving AI | Department of Computer Science and Technology
At a time when there's widespread interest in self-improving AI, researchers here have tackled one of the central challenges affecting its development by using an approach adapted from evolutionary bi...