多任务强化学习 Multi-Task RL

Gradient

Modularization

Task Representation

  • CARE: Multi-Task Reinforcement Learning with Context-based Representations (ICML 2021)

  • Multi-task Reinforcement Learning with Task Representation Method (ICLR2022 Workshop)

Curriculum Learning

  • CAMRL: Curriculum-based Asymmetric Multi-task Reinforcement Learning (TPAMI 2022)

Offline & Transformer

  • GATO: A Generalist Agent (DeepMind 2022)

  • MGDT: Multi-Game Decision Transformers (Google Research 2022)

  • AD: In-context Reinforcement Learning with Algorithm Distillation (DeepMind 2022)

  • Uni[MASK]: Unified Inference in Sequential Decision Problems (NeurIPS 2022)

多目标强化学习 Multi-Goal RL

  • UVFA: Universal Value Function Approximators (ICML 2015)

  • HER: Hindsight Experience Replay (NeurIPS 2017)

  • GoalGAN: Automatic Goal Generation for Reinforcement Learning Agents (ICML 2018)

  • VisualHER: Addressing Sample Complexity in Visual Tasks Using HER and Hallucinatory GANs (NeurIPS 2019)

  • CURIOUS: Intrinsically Motivated Modular Multi-goal Reinforcement Learning (ICML 2019)

Last modification：December 25, 2022
© Allow specification reprint
如果觉得我的文章对你有用，请随意赞赏