Review request:
As usual for the time of year, I'll be looking for #IROS2026 reviewers. Highly interesting stack of papers on #ReinforcementLearning #Sim2Real #RewardLearning #LLMs #DataEfficiency #RoboticManipulation

Reach out with your ID or papercept registered mail address and background.

A new T-RO paper introduces HARMONIOUS, a real-time control system enabling humanoid robots to play board games while safely interacting with humans through multi-modal sensing (touch, proximity, vision).
ieeexplore.ieee.org/document/107...

#HumanRobotInteraction #RoboticManipulation

Researchers unveiled CroSTAta, a transformer that improves robotic manipulation, delivering over two‑fold accuracy gains on precision tasks versus standard cross‑attention. Code on GitHub. https://getnews.me/crostata-transformer-boosts-robotic-manipulation-accuracy/ #crostata #roboticmanipulation

ReLAM predicts keypoint‑based subgoals to generate dense rewards for robotic manipulation; the September 2025 paper reports faster learning and success rates than prior methods. https://getnews.me/relam-boosts-visual-robotic-manipulation-with-learned-reward-anticipation/ #relam #roboticmanipulation

GP3 lets robots infer depth and camera parameters from multiple RGB views, creating a compact 3‑D scene representation without depth sensors. Paper posted 19 Sep 2025. https://getnews.me/gp3-introduces-3-d-geometry-aware-policy-for-robotic-manipulation/ #gp3 #roboticmanipulation

Geometric Red‑Teaming creates CrashShapes that cut task success from ~90% to 22.5%, while blue‑teaming restores up to 60 percentage points of performance. Read more: https://getnews.me/geometric-red-teaming-boosts-robustness-in-robotic-manipulation/ #geometricredteam #roboticmanipulation #corl2025

Robo-Dopamine: General Process Reward Modeling for High-Precision Robotic Manipulation
Cheng Chi, Guocai Yao et al.
Paper
Details
#Robotics #ReinforcementLearning #RoboticManipulation