D4RL
PulseAugur coverage of D4RL — every cluster mentioning D4RL across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New Bayesian framework enhances offline multi-agent reinforcement learning
Researchers have developed a Bayesian framework for offline multi-agent reinforcement learning (MARL) that addresses challenges posed by limited dataset coverage and potential out-of-distribution errors. The proposed me…
-
DT-JEPA framework learns generalizable knowledge for offline reinforcement learning
Researchers have introduced DT-JEPA, a novel framework for offline reinforcement learning that combines Decision Transformer (DT) with Joint Embedding Predictive Architectures (JEPA). This approach aims to learn general…
-
New VRD Framework Tackles Out-of-Distribution Mode Averaging in Offline RL
Researchers have developed Value-Rectified Distillation (VRD), a new framework to address the challenge of Out-of-Distribution (OOD) mode averaging in offline Reinforcement Learning (RL). This problem occurs when distil…
-
New research explores higher-order action supervision and action shaping for robust RL policies
Two new arXiv papers explore advanced techniques in reinforcement learning (RL) and imitation learning (IL) to improve policy robustness and control. The first paper introduces a method for higher-order action supervisi…
-
New imitation learning methods enhance robot reliability and performance · 4 sources tracked
Researchers are developing new methods to improve imitation learning for robots, focusing on enhancing reliability and performance with imperfect data. Rewind-IL introduces a framework for online failure detection and r…
-
New Trust Guided Decision Transformer improves AI long-term decision-making
Researchers have developed a new method called Trust Guided Decision Transformer (TGDT) to improve the performance of Decision Transformers in long-term AI decision-making. Standard Decision Transformers struggle when t…
-
New method boosts efficiency in learning AI reward models from human preferences
Researchers have developed PreferenceEKF, a novel method for active learning in reinforcement learning from human feedback (RLHF). This approach addresses the sample inefficiency of RLHF by framing active preference lea…
-
New STITCH-OPE framework uses guided diffusion for off-policy evaluation
Researchers have developed STITCH-OPE, a novel framework for off-policy evaluation (OPE) that utilizes guided diffusion models. This method is designed to handle high-dimensional, long-horizon problems in fields like ro…
-
New SMAC method enables robust offline-to-online reinforcement learning transfer
Researchers have developed a new method called Score-Matched Actor-Critic (SMAC) to improve the transfer of reinforcement learning models from offline to online environments. Traditional methods often see performance dr…
-
New RS-Diffuser framework offers risk-sensitive offline reinforcement learning
Researchers have developed RS-Diffuser, a novel framework for risk-sensitive offline reinforcement learning. This approach combines diffusion-based trajectory generation with distributional value critics to allow for fl…
-
New ROMI method advances offline reinforcement learning, outperforming prior models
Researchers have introduced ROMI, a novel method for model-based offline reinforcement learning that addresses key challenges in adversarial model learning. Unlike previous approaches like RAMBO, which struggled with co…
-
New research enhances diffusion models for robust RL and safe planning
Researchers are developing new methods to improve the robustness and safety of diffusion models in reinforcement learning and planning tasks. One approach, Robust Regularized Policy Iteration (RRPI), addresses transitio…
-
New MPDiffuser framework enhances diffusion model control for robotics
Researchers have developed a new framework called Model Predictive Diffuser (MPDiffuser) to improve the reliability of diffusion models in offline decision-making tasks. This approach combines a diffusion planner with a…
-
New research explores Q-learning stability and offline RL methods
Two new research papers explore advancements in reinforcement learning techniques. One paper introduces Drift Q-Learning, a method that combines a drift-based behavioral regularizer with critic-driven policy improvement…
-
New MoMa QL framework boosts RL efficiency with moment matching
Researchers have introduced Moment Matching Q-Learning (MoMa QL), a novel framework designed to address the inference latency issues in score-based and flow-based generative models used in reinforcement learning. MoMa Q…
-
New SPAR framework improves offline policy improvement in AI
Researchers have introduced Support-Preserving Action Rectification (SPAR), a novel framework designed to address the inherent conflict in offline policy improvement. SPAR reframes global learning as a local residual re…
-
New research advances policy optimization for robotics and LLMs
Researchers have introduced several new methods to enhance policy optimization in reinforcement learning, particularly for complex tasks involving robotics and large language models. MODIP aims to efficiently fine-tune …
-
New COOPO framework boosts reinforcement learning efficiency
Researchers have developed a new framework called COOPO (Cyclic Offline-Online Policy Optimization) to address limitations in offline and online reinforcement learning. This method repeatedly cycles between offline trai…
-
SlimDT paper proposes injecting RTG outside sequential modeling
Researchers have developed SlimDT, a modification of the Decision Transformer (DT) model for offline reinforcement learning. SlimDT removes the Return-to-Go (RTG) token from the autoregressive sequence, instead injectin…