Pipette: An Embodied Simulation Platform, Benchmark, and Data-Efficient Augmentation Framework for Wet-Lab Robotics

arXiv cs.AI · 6 hours ago

Researchers released Pipette, a simulation platform and benchmark for training wet-lab robots with data-efficient augmentation techniques for biomedical experiments. The platform includes over 100 open-source laboratory assets, supports three robotic arms, and a data augmentation pipeline that improves model performance from 40.4% to 71.8% success rate using only 30 demonstrations per task. This enables faster training of robot learning systems for laboratory automation with limited manual data.