Adroit

The 'Adroit' dataset is a benchmark that contains a collection of robotic manipulation tasks used to evaluate planning-based reinforcement learning algorithms in continuous control environments.

Papers using Adroit (10)

RIZE: Adaptive Regularization for Imitation Learning2025 · 1 citesXQCfD: Accelerating Fast Actor-Critic Algorithms with Prior Data and Prior Policies2026In-Context Planning with Latent Temporal Abstractions2026In-Context Compositional Q-Learning for Offline Reinforcement Learning2025Online Pre-Training for Offline-to-Online Reinforcement Learning2025ADG: Ambient Diffusion-Guided Dataset Recovery for Corruption-Robust Offline Reinforcement Learning2025AlignIQL: Policy Alignment in Implicit Q-Learning through Constrained Optimization2024Tackling Data Corruption in Offline Reinforcement Learning via Sequence Modeling2024Energy-Guided Diffusion Sampling for Offline-to-Online Reinforcement Learning2024Task-Agnostic Pre-training and Task-Guided Fine-tuning for Versatile Diffusion Planner2024