Provably Safe Offline-to-Online RL: Decoupling Learning from Data-Driven Safety Enforcement.

Kaitong Cai, Jusheng Zhang, Keze Wang

Browse the full ACL paper archive.