Researchers from Nanjing University and UC Berkeley have unveiled a clustering-based reinforcement learning framework that balances novelty and ...