Action Space

Action space, in reinforcement learning, refers to the set of all possible actions an agent can take within an environment. Current research focuses on efficiently handling large or complex action spaces, particularly in multi-agent systems and continuous control problems, employing techniques like action discretization, factorization, and the use of large language models for guidance. These advancements are crucial for scaling reinforcement learning to real-world applications, such as robotics and resource management, where high-dimensional and nuanced action choices are common. Improved methods for handling action spaces directly impact the sample efficiency and overall performance of reinforcement learning algorithms.

Papers

November 17, 2024

An Investigation of Offline Reinforcement Learning in Factorisable Action Spaces
Alex Beeson, David Ireland, Giovanni Montana
Offline Reinforcement Learning Action Space Comprehensive Investigation

November 4, 2024

Benchmarking Vision, Language, & Action Models on Robotic Learning Tasks
Pranav Guruprasad, Harshvardhan Sikka, Jaewoo Song, Yangyue Wang, Paul Pu Liang
Human Language Action Space Robot Learning Action Model Action Generation Vision Language Action General Purpose Vision Benchmark Robot Task

October 28, 2024

October 25, 2024

October 22, 2024

Composing Diffusion Policies for Few-shot Learning of Movement Trajectories
Omkar Patil, Anant Sah, Nakul Gopalan
Shot Learning Action Space Handwritten Trajectory Movement Related Information Policy Diffusion

October 21, 2024

Diffusion Transformer Policy
Zhi Hou, Tianyi Zhang, Yuwen Xiong, Hengjun Pu, Chengyang Zhao, Ronglei Tong, Yu Qiao, Jifeng Dai, Yuntao Chen
Action Space Diffusion Transformer Robotic Datasets Vision Language Action

October 18, 2024

October 17, 2024

October 15, 2024

Learning Agents With Prioritization and Parameter Noise in Continuous State and Action Space
Rajesh Mangannavar, Gopalakrishnan Srinivasaraghavan
Action Space Deep Q Learning Deep Deterministic Policy Gradient Continuous Speech Learning Agent Priority Based Noise Modeling Robust Deep Reinforcement Learning

October 12, 2024

Reinforcement Learning in Hyperbolic Spaces: Models and Experiments
Vladimir Jaćimović, Zinaid Kapić, Aladin Crnkić
Reinforcement Learning Full Model Action Space Optical Experiment Black Box Optimization Unknown Environment Hyperbolic Space Dynamical Model

October 8, 2024

October 5, 2024

On the Sample Complexity of a Policy Gradient Algorithm with Occupancy Approximation for General Utility Reinforcement Learning
Anas Barakat, Souradip Chakraborty, Peihong Yu, Pratap Tokekar, Amrit Singh Bedi
Reinforcement Learning Sample Complexity Action Space Policy OpTimization Policy Gradient Algorithm Utility Learning

Action Space

Papers

An Investigation of Offline Reinforcement Learning in Factorisable Action Spaces

Benchmarking Vision, Language, & Action Models on Robotic Learning Tasks

Learning State Conditioned Linear Mappings for Low-Dimensional Control of Robotic Manipulators

Offline Reinforcement Learning With Combinatorial Action Spaces

SAD: State-Action Distillation for In-Context Reinforcement Learning under Random Policies

An Enhanced Hierarchical Planning Framework for Multi-Robot Autonomous Exploration

Composing Diffusion Policies for Few-shot Learning of Movement Trajectories

Diffusion Transformer Policy

Online Reinforcement Learning with Passive Memory

Transfer Reinforcement Learning in Heterogeneous Action Spaces using Subgoal Mapping

RecoveryChaining: Learning Local Recovery Policies for Robust Manipulation

AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents

Learning Agents With Prioritization and Parameter Noise in Continuous State and Action Space

Reinforcement Learning in Hyperbolic Spaces: Models and Experiments

Towards Synergistic, Generalized, and Efficient Dual-System for Robotic Manipulation

Offline Hierarchical Reinforcement Learning via Inverse Optimization

Imitation Learning with Limited Actions via Diffusion Planners and Deep Koopman Controllers

Learning in complex action spaces without policy gradients

ConceptAgent: LLM-Driven Precondition Grounding and Tree Search for Robust Task Planning and Execution

On the Sample Complexity of a Policy Gradient Algorithm with Occupancy Approximation for General Utility Reinforcement Learning