MuJoCo
PulseAugur coverage of MuJoCo — every cluster mentioning MuJoCo across labs, papers, and developer communities, ranked by signal.
11 day(s) with sentiment data
-
TensorWorld raises $100M+ for physics-driven AI models
TensorWorld, a four-month-old startup, has secured over $100 million in funding to develop AI models that understand and predict physical world interactions. The company focuses on a physics-driven approach, learning ph…
-
MuJoCo README translated to Chinese to simplify setup
A Chinese translation of the MuJoCo README has been published, aiming to simplify the setup process for Chinese-speaking users. The translated repository includes a quick-start guide that reorders and clarifies the orig…
-
New Ordered Policy Optimization method enhances reinforcement learning stability
Researchers have introduced Ordered Policy Optimization (OPO), a novel method designed to improve the stability and effectiveness of policy updates in reinforcement learning. Unlike traditional Proximal Policy Optimizat…
-
New system grounds VLM spatial reasoning in physics engine feedback
Researchers have developed Spatial Harness, a novel system designed to enhance the spatial reasoning capabilities of vision-language models (VLMs). This system grounds VLM reasoning in explicit, executable world states …
-
New $\nabla$Mimic method drastically cuts humanoid robot control training time
Researchers have developed a new method called $\nabla$Mimic that significantly improves the efficiency of humanoid robot control by utilizing differentiable simulation. This approach leverages exact gradients from the …
-
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…
-
Patchwork system slashes simulation resource use by 92%
Researchers have developed Patchwork, a novel system designed to optimize resource sharing in domain-randomized embodied simulations. This system separates common scene resources into shared templates and specific mutat…
-
New PINN model learns hybrid systems with discrete jumps
Researchers have developed Hybrid-time PINNs (HT-PINNs), a novel approach to modeling hybrid dynamical systems that exhibit both continuous flow and discrete jumps. Unlike traditional PINNs that struggle with discontinu…
-
New regularized RL algorithm enhances sim-to-real transfer for robots
A new reinforcement learning algorithm, Soft MellowMax Actor-Critic (SMM-AC), has been developed to improve the transfer of policies from simulation to real-world robots. This algorithm, based on the soft mellowmax valu…
-
New pipeline generates thousands of robot demonstrations from single human video
Researchers have developed Real2Sem2Real (R2Sem2R), a novel pipeline that transforms a single human RGB video demonstration into thousands of robot demonstrations. This method utilizes DINOv3-distilled Gaussian splats t…
-
Two new arXiv papers explore generative models for RL and industrial ranking
Two new research papers were published on arXiv, both focusing on advancements in generative models. The first, "Doob-Based Generative Policy Optimization," introduces a novel framework for reinforcement learning that a…
-
New framework verifies 3D scene reconstructions for robot policy evaluation
Researchers have developed YUBI-Real2Sim, a framework designed to improve the reliability of evaluating robotic policies in simulation by verifying 3D Gaussian Splatting (3DGS) scene reconstructions. The system addresse…
-
Generative Trajectory Planning (GTP) framework enhances AI decision-making
Researchers have introduced Generative Trajectory Planning (GTP), a novel model-based framework designed to improve online decision-making by learning a generative model as a proposal distribution for action sequences. …
-
New framework improves drone landings on ships in rough seas
Researchers have developed a new control framework for autonomous drone landings on moving ships in rough seas. This framework uses a pretrained model to explicitly predict the ship's attitude, which helps in selecting …
-
Google releases Gemini 4 Argon, its most capable frontier model yet · 10 sources tracked
Google DeepMind has launched Gemini 4 Argon, its latest frontier model, emphasizing its advanced capabilities in complex tasks such as coding, cybersecurity, and scientific research. The model is designed with robust sa…
-
New Geometric Approach Enhances Locomotion Learning in Reinforcement Agents
Researchers have developed GeZo-SAC, a novel approach to Soft Actor-Critic methods that enhances locomotion learning by adapting critic pessimism using geometric representations. This method utilizes zonotopes to repres…
-
New framework links robot scene reconstruction and policy development for real-world tasks
Researchers have developed Agentic RSR, a framework that integrates scene reconstruction, policy development, and real-robot execution for manipulation tasks. This system takes workspace videos and task descriptions to …
-
New AI method generates simulation-ready collision geometry from images
Researchers have developed I2CD, a novel method for generating convex collision geometry directly from single RGB images. This approach bypasses the traditional, time-consuming pipeline of reconstructing 3D models and t…
-
autotrust releases Gemma-4 based decision model in bf16 and NVFP4 quantized versions
The autotrust/GEV-26B-Decide model, based on Gemma-4, has been released in two versions: a standard bf16 model and a quantized NVFP4 version optimized for vLLM. The NVFP4 version significantly reduces memory requirement…
-
New ROS2 Framework Enhances Real-Time Control for Multi-Robot Systems
Researchers have developed multipanda_ros2, an open-source ROS2 framework designed for real-time control of multiple Franka Robotics arms. This architecture aims to bridge the simulation-to-reality gap by maintaining a …