ICML 2020 accepted papers on Reinforcement learning
103 of the 1,087 papers accepted at ICML 2020 (9.5%) are on Reinforcement learning, compared with 9.9% at ICML 2019. They fall into 6 topics.
See ICML 2020 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 · 38 papers
- Revisiting Fundamentals of Experience Replay (79 citations)
- A Game Theoretic Framework for Model Based Reinforcement Learning (53 citations)
- Controlling Overestimation Bias with Truncated Mixture of Continuous Distributional Quantile Critics (50 citations)
- Responsive Safety in Reinforcement Learning by PID Lagrangian Methods (43 citations)
- Monte-Carlo Tree Search as Regularized Policy Optimization (38 citations)
- A distributional view on multi-objective policy optimization (31 citations)
- Ready Policy One: World Building Through Active Learning (27 citations)
- Discount Factor as a Regularizer in Reinforcement Learning (23 citations)
- Minimax Weight and Q-Function Learning for Off-Policy Evaluation (23 citations)
- Representations for Stable Off-Policy Reinforcement Learning (13 citations)
- Striving for Simplicity and Performance in Off-Policy DRL: Output Normalization and Non-Uniform Sampling (13 citations)
- Bidirectional Model-based Policy Optimization (12 citations)
and 26 more in the interactive map.
Visual and representation RL · 37 papers
- Agent57: Outperforming the Atari Human Benchmark (226 citations)
- CURL: Contrastive Unsupervised Representations for Reinforcement Learning (215 citations)
- Leveraging Procedural Generation to Benchmark Reinforcement Learning (163 citations)
- Reinforcement Learning for Integer Programming: Learning to Cut (73 citations)
- Bootstrap Latent-Predictive Representations for Multitask Reinforcement Learning (58 citations)
- Skew-Fit: State-Covering Self-Supervised Reinforcement Learning (58 citations)
- Learning to Navigate The Synthetically Accessible Chemical Space Using Reinforcement Learning (51 citations)
- One Policy to Control Them All: Shared Modular Policies for Agent-Agnostic Control (43 citations)
- Maximum Entropy Gain Exploration for Long Horizon Multi-goal Reinforcement Learning (38 citations)
- Invariant Causal Prediction for Block MDPs (36 citations)
- Adaptive Reward-Poisoning Attacks against Reinforcement Learning (33 citations)
- Context-aware Dynamics Model for Generalization in Model-Based Reinforcement Learning (25 citations)
and 25 more in the interactive map.
MDPs and planning · 20 papers
- Provably Efficient Exploration in Policy Optimization (80 citations)
- Safe Reinforcement Learning in Constrained Markov Decision Processes (52 citations)
- Kinematic State Abstraction and Provably Efficient Rich-Observation Reinforcement Learning (45 citations)
- Reward-Free Exploration for Reinforcement Learning (41 citations)
- Policy Teaching via Environment Poisoning: Training-time Adversarial Attacks against Reinforcement Learning (36 citations)
- Efficiently Solving MDPs with Stochastic Mirror Descent (17 citations)
- Optimizing for the Future in Non-Stationary MDPs (17 citations)
- Constrained Markov Decision Processes via Backward Value Functions (16 citations)
- A Finite-Time Analysis of Q-Learning with Neural Network Function Approximation (15 citations)
- Minimax-Optimal Off-Policy Evaluation with Linear Function Approximation (15 citations)
- A Markov Decision Process Model for Socio-Economic Systems Impacted by Climate Change (14 citations)
- Provable Representation Learning for Imitation Learning via Bi-level Optimization (14 citations)
and 8 more in the interactive map.
Imitation learning · 5 papers
- Safe Imitation Learning via Fast Bayesian Reward Inference from Preferences (41 citations)
- Intrinsic Reward Driven Imitation Learning via Generative Model (12 citations)
- Domain Adaptive Imitation Learning (11 citations)
- Variational Imitation Learning with Diverse-quality Demonstrations (3 citations)
- Deep PQR: Solving Inverse Reinforcement Learning using Anchor Actions (2 citations)
