| 1 |
The Impact of Quantization and Pruning on Deep Reinforcement Learning Models |
研究量化和剪枝对深度强化学习模型性能的影响，旨在资源受限环境下的高效部署。 |
reinforcement learning deep reinforcement learning DRL |
|
|
| 2 |
Tackling Data Corruption in Offline Reinforcement Learning via Sequence Modeling |
提出RDT，通过序列建模解决离线强化学习中的数据损坏问题 |
reinforcement learning offline RL offline reinforcement learning |
✅ |
|
| 3 |
Hindsight Preference Learning for Offline Preference-based Reinforcement Learning |
提出HPL：利用后见之明偏好学习解决离线偏好强化学习中的信用分配问题 |
reinforcement learning preference learning |
✅ |
|
| 4 |
Graph Reinforcement Learning for Power Grids: A Comprehensive Survey |
图强化学习用于电力系统控制：综述电力网络中基于图强化学习的控制方法。 |
reinforcement learning representation learning |
|
|
| 5 |
Improving Knowledge Distillation in Transfer Learning with Layer-wise Learning Rates |
提出层级学习率的知识蒸馏迁移学习方法，提升复杂任务性能。 |
distillation |
|
|
| 6 |
Explorative Imitation Learning: A Path Signature Approach for Continuous Environments |
提出基于路径签名和探索的模仿学习方法CILO，用于连续控制环境。 |
imitation learning |
|
|
| 7 |
Understanding the Gains from Repeated Self-Distillation |
研究重复自蒸馏的增益，揭示其在降低线性回归风险方面的潜力 |
distillation |
|
|
| 8 |
Using Petri Nets as an Integrated Constraint Mechanism for Reinforcement Learning Tasks |
提出基于Petri网的强化学习约束框架，提升AI可信度并应用于交通信号控制 |
reinforcement learning |
|
|
| 9 |
Simplifying Deep Temporal Difference Learning |
提出PQN：一种简化的深度在线Q学习算法，无需目标网络和经验回放，且性能优异。 |
reinforcement learning PPO |
|
|