robot control
PulseAugur coverage of robot control — every cluster mentioning robot control across labs, papers, and developer communities, ranked by signal.
-
Second-Order Drifting Models accelerate generative AI training dynamics
Researchers have introduced Second-Order Drifting Models, an advancement in one-step generative models that evolve distributions during training. By incorporating artificial velocity variables into generated samples, th…
-
New DUPO method improves reinforcement learning with diffusion models
Researchers have introduced Diffusion-Guided Uncertainty-Aware Delayed Policy Optimization (DUPO), a novel approach to address performance degradation in reinforcement learning caused by delayed feedback. DUPO explicitl…
-
New framework enables robots to adapt to new environments without retraining
Researchers have introduced In-Context World Modeling (ICWM), a new framework designed to improve the adaptability of robotic policies. ICWM treats system identification as an in-context adaptation problem, enabling rob…
-
Image editing models replace video generation in robot control systems
Researchers have developed ImageWAM, a novel framework that utilizes pretrained image editing models for robot control, challenging the necessity of video generation in World Action Models (WAMs). This approach signific…
-
ETH Zurich Releases Comprehensive Robotics Course Covering Foundation Models
ETH Zurich has released a popular robotics course that covers topics from foundational robotics principles to foundation models. The curriculum includes robot control, learning-based robotics, and the latest in foundati…
-
RepWAM model enhances robot manipulation with visual-action tokenization
Researchers have introduced RepWAM, a novel world action model designed for robot manipulation. This model utilizes semantic visual-action tokenization to create a latent space that better connects language instructions…
-
ModuLoop framework uses LLMs for robotic control code generation
Researchers have developed a new framework called ModuLoop to generate low-level code for robotic control tasks. This system utilizes a pre-trained Large Language Model (LLM) to plan and create code, which is then execu…