Researchers from ETH Zurich and UC Berkeley Introduce MaxInfoRL: A New Reinforcement Learning Framework for Balancing Intrinsic and Extrinsic Exploration – MarkTechPost