ICLR 2024 accepted papers on Reinforcement learning
187 of the 2,259 papers accepted at ICLR 2024 (8.3%) are on Reinforcement learning, compared with 9.4% at ICLR 2023. They fall into 6 topics.
See ICLR 2024 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Visual and representation RL · 74 papers
- TorchRL: A data-driven decision-making library for PyTorch (16 citations)
- Predictive auxiliary objectives in deep RL mimic learning in the brain (5 citations)
- RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches (5 citations)
- Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Tasks (4 citations)
- Universal Humanoid Motion Representations for Physics-Based Control (4 citations)
- DrM: Mastering Visual Reinforcement Learning through Dormant Ratio Minimization (3 citations)
- Jumanji: a Diverse Suite of Scalable Reinforcement Learning Environments in JAX (3 citations)
- Multi-Task Reinforcement Learning with Mixture of Orthogonal Experts (3 citations)
- Proximal Policy Gradient Arborescence for Quality Diversity Reinforcement Learning (3 citations)
- Revisiting Data Augmentation in Deep Reinforcement Learning (3 citations)
- Searching for High-Value Molecules Using Reinforcement Learning and Transformers (3 citations)
- AMAGO: Scalable In-Context Reinforcement Learning for Adaptive Agents (2 citations)
and 62 more in the interactive map.
Offline RL · 29 papers
- Unleashing the Power of Pre-trained Language Models for Offline Reinforcement Learning (5 citations)
- When should we prefer Decision Transformers for Offline Reinforcement Learning? (5 citations)
- Safe Offline Reinforcement Learning with Feasibility-Guided Diffusion Model (3 citations)
- Consistency Models as a Rich and Efficient Policy Class for Reinforcement Learning (2 citations)
- Delphic Offline Reinforcement Learning under Nonidentifiable Hidden Confounding (2 citations)
- Reasoning with Latent Diffusion in Offline Reinforcement Learning (2 citations)
- Score Regularized Policy Optimization through Diffusion Behavior (2 citations)
- The Generalization Gap in Offline Reinforcement Learning (2 citations)
- Uni-O4: Unifying Online and Offline Deep Reinforcement Learning with Multi-Step On-Policy Optimization (2 citations)
- Improving Offline RL by Blending Heuristics (1 citation)
- Transformers as Decision Makers: Provable In-Context Reinforcement Learning via Supervised Pretraining (1 citation)
- Compositional Conservatism: A Transductive Approach in Offline Reinforcement Learning
and 17 more in the interactive map.
Policy gradient and actor-critic · 28 papers
- TD-MPC2: Scalable, Robust World Models for Continuous Control (8 citations)
- S$2$AC: Energy-Based Reinforcement Learning with Stein Soft Actor Critic (2 citations)
- Fast Value Tracking for Deep Reinforcement Learning (1 citation)
- Identifying Policy Gradient Subspaces (1 citation)
- Kernel Metric Learning for In-Sample Off-Policy Evaluation of Deterministic RL Policies (1 citation)
- Physics-Regulated Deep Reinforcement Learning: Invariant Embeddings (1 citation)
- Adaptive Regularization of Representation Rank as an Implicit Constraint of Bellman Equation
- Decoupling regularization from the action space
- Diverse Projection Ensembles for Distributional Reinforcement Learning
- Game-Theoretic Robust Reinforcement Learning Handles Temporally-Coupled Perturbations
- Impact of Computation in Integral Reinforcement Learning for Continuous-Time Control
- Maximum Entropy Model Correction in Reinforcement Learning
and 16 more in the interactive map.
MDPs and planning · 21 papers
- Provable and Practical: Efficient Exploration in Reinforcement Learning via Langevin Monte Carlo (5 citations)
- Achieving Sample and Computational Efficient Reinforcement Learning by Action Space Reduction via Grouping (1 citation)
- Beyond Stationarity: Convergence Analysis of Stochastic Softmax Policy Gradient Methods (1 citation)
- Bridging State and History Representations: Understanding Self-Predictive RL (1 citation)
- Finite-Time Analysis of On-Policy Heterogeneous Federated Reinforcement Learning (1 citation)
- Goodhart's Law in Reinforcement Learning (1 citation)
- Sample-Efficient Learning of POMDPs with Multiple Observations In Hindsight (1 citation)
- Soft Robust MDPs and Risk-Sensitive MDPs: Equivalence, Policy Gradient, and Sample Complexity (1 citation)
- Tree Search-Based Policy Optimization under Stochastic Execution Delay (1 citation)
- A Policy Gradient Method for Confounded POMDPs
- Demonstration-Regularized RL
- Provable Benefits of Multi-task RL under Non-Markovian Decision Making Processes
and 9 more in the interactive map.
RL from feedback · 19 papers
- Text2Reward: Reward Shaping with Language Models for Reinforcement Learning (9 citations)
- Duolando: Follower GPT with Off-Policy Reinforcement Learning for Dance Accompaniment (7 citations)
- Statistical Rejection Sampling Improves Preference Optimization (7 citations)
- Uni-RLHF: Universal Platform and Benchmark Suite for Reinforcement Learning with Diverse Human Feedback (3 citations)
- Zeroth-Order Optimization Meets Human Feedback: Provable Learning via Ranking Oracles (3 citations)
- Reward Model Ensembles Help Mitigate Overoptimization (2 citations)
- AlignDiff: Aligning Diverse Human Preferences via Behavior-Customisable Diffusion Model (1 citation)
- Confronting Reward Model Overoptimization with Constrained RLHF (1 citation)
- PARL: A Unified Framework for Policy Alignment in Reinforcement Learning from Human Feedback (1 citation)
- $\mathcal{B}$-Coder: Value-Based Deep Reinforcement Learning for Program Synthesis
- Hindsight PRIORs for Reward Learning from Human Preferences
- Improving Generalization of Alignment with Human Preferences through Group Invariant Learning
and 7 more in the interactive map.
Imitation learning · 16 papers
- RLIF: Interactive Imitation Learning as Reinforcement Learning (7 citations)
- Blending Imitation and Reinforcement Learning for Robust Policy Improvement (2 citations)
- Fourier Transporter: Bi-Equivariant Robotic Manipulation in 3D (2 citations)
- Massively Scalable Inverse Reinforcement Learning in Google Maps (2 citations)
- Memory-Consistent Neural Networks for Imitation Learning (1 citation)
- SEABO: A Simple Search-Based Method for Offline Imitation Learning (1 citation)
- Towards Diverse Behaviors: A Benchmark for Imitation Learning with Human Demonstrations (1 citation)
- Adversarial Imitation Learning via Boosting
- CCIL: Continuity-Based Data Augmentation for Corrective Imitation Learning
- Dual RL: Unification and New Methods for Reinforcement and Imitation Learning
- Learning to Act without Actions
- Quantifying the Sensitivity of Inverse Reinforcement Learning to Misspecification
and 4 more in the interactive map.