ICML 2017 accepted papers on Reinforcement learning
25 of the 435 papers accepted at ICML 2017 (5.7%) are on Reinforcement learning, compared with 5.3% at ICML 2016. They fall into 5 topics.
See ICML 2017 in the interactive map
Submitting to ICML? Checklist →
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Policy gradient and actor-critic · 11 papers
- Reinforcement Learning with Deep Energy-Based Policies (409 citations)
- A Distributional Perspective on Reinforcement Learning (238 citations)
- Averaged-DQN: Variance Reduction and Stabilization for Deep Reinforcement Learning (154 citations)
- Constrained Policy Optimization (109 citations)
- Improving Stochastic Policy Gradients in Continuous Control with Deep Reinforcement Learning using the Beta Distribution (99 citations)
- Combining Model-Based and Model-Free Updates for Trajectory-Centric Reinforcement Learning (86 citations)
- Consistent On-Line Off-Policy Evaluation (32 citations)
- An Alternative Softmax Operator for Reinforcement Learning (25 citations)
- Efficient Nonmyopic Active Search (23 citations)
- Data-Efficient Policy Evaluation Through Behavior Policy Search (13 citations)
- Spectral Learning from a Single Trajectory under Finite-State Policies (2 citations)
Visual and representation RL · 8 papers
- Robust Adversarial Reinforcement Learning (366 citations)
- FeUdal Networks for Hierarchical Reinforcement Learning (245 citations)
- Modular Multitask Reinforcement Learning with Policy Sketches (214 citations)
- DARLA: Improving Zero-Shot Transfer in Reinforcement Learning (127 citations)
- Fake News Mitigation via Point Process Based Intervention (114 citations)
- Zero-Shot Task Generalization with Multi-Task Deep Reinforcement Learning (109 citations)
- A Laplacian Framework for Option Discovery in Reinforcement Learning (73 citations)
- Hierarchy Through Composition with Multitask LMDPs (18 citations)
MDPs and planning · 3 papers
- Contextual Decision Processes with low Bellman rank are PAC-Learnable (143 citations)
- Unifying Task Specification in Reinforcement Learning (17 citations)
- Fairness in Reinforcement Learning (7 citations)
