Speakers

Ron Alterovitz

University of North Carolina, USA

Talk title: Accelerating Scientific Discovery Through Laboratory Automation: Mobile Robots and Experiment Orchestration Software

Abstract: Despite producing cutting-edge discoveries, most modern laboratories in chemistry, biochemistry, and materials science rely heavily on manual processes that limit experimental throughput and reproducibility. Robotics and automation offer transformative potential to conduct experiments faster, more safely, and with greater precision, enabling scientists to address societal challenges on accelerated timescales. In this talk, we define five levels of laboratory automation, from laboratory assistance to full automation, enabling us to classify laboratories and track progress in automation. We also present two key enabling technologies for higher levels of automation. First, I will demonstrate how mobile manipulation robots can automate tasks in laboratories where fixed automation is impractical. Specifically, I will present a method enabling mobile manipulation robots to transport samples in syringes over many meters and then perform precise injection tasks requiring millimeter-level accuracy. Second, I will present the Experiment Orchestration System (EOS), an open-source framework providing a comprehensive laboratory automation software infrastructure. EOS enables users to define labs, devices, and experiments using YAML and Python plugins, and EOS then manages and executes the automation. The system implements autonomous campaigns, parameter optimization, and result aggregation, significantly reducing automation implementation barriers. Together, these advances can ease the creation of new automated laboratories that accelerate scientific discovery.

Bio: Ron Alterovitz is the Lawrence Grossberg Distinguished Professor in the Department of Computer Science at the University of North Carolina at Chapel Hill. He leads the Computational Robotics Research Group, which focuses on increasing the autonomy of robots by developing novel algorithms for robots to learn and plan their motions. Prior to joining UNC-Chapel Hill in 2009, Dr. Alterovitz earned his B.S. with Honors from the California Institute of Technology (Caltech), completed his Ph.D. at the University of California, Berkeley, and conducted postdoctoral research at the UCSF Comprehensive Cancer Center and a French National Center for Scientific Research (CNRS) lab in Toulouse, France. Dr. Alterovitz is co-inventor on three patents and has received multiple best paper awards at robotics and computer-assisted medicine conferences including RSS, ICRA, IROS, and MIUA. He is the recipient of an NIH Ruth L. Kirschstein National Research Service Award, two UNC Computer Science Department Excellence in Teaching Awards, an NSF Early Career Development (CAREER) Award, and the Presidential Early Career Award for Scientists and Engineers (PECASE), the highest honor bestowed by the United States Government on science and engineering professionals in the early stages of their independent research careers. In 2025, he became a Fellow of the IEEE (the world’s largest technical professional society).

Glen Berseth

Talk title: From Tokens to Actions: How to Build General Robot Policies

Abstract: Using large models to create robot control solutions across a wide distribution of tasks has been accelerating the success of robotics in the real world. This tutorial will show how to build a Generalist Robotics Policy from scratch. Generalist Robotics Policies are large models for robotics that are trained using large amounts of interaction data. We reimplement a set of recent GRP models step by step starting from a few lines of transformer code and show training the model over data from the open-x embodiment dataset. We also discuss the data, infrastructure, methods and limitations of these methods.

Bio: Glen Berseth is an assistant professor at the Université de Montréal, a core academic member of Mila, Canada CIFAR AI chair, member L’Institut Courtois, and co-director of the Robotics and Embodied AI Lab (REAL). He was a postdoctoral researcher with Berkeley Artificial Intelligence Research (BAIR) working in the Robotic AI & Learning (RAIL) lab with Sergey Levine. Glen completed his NSERC-supported Ph.D. in Computer Science at the University of British Columbia in 2019, where he worked with Michiel van de Panne. His current research focuses on machine learning and solving real-world sequential decision-making problems (planning/RL), such as robotics, scientific discovery and adaptive clean technology. The specifics of his research have covered the areas of human-robot collaboration, generalization, reinforcement learning, continual learning, meta-learning, multi-agent learning, and hierarchical learning. Dr. Berseth has published across the top venues in machine learning, robotics, and computer animation in his work. He also teaches courses on data science and robot learning at Université de Montréal and Mila, covering the most recent research on machine learning techniques for creating generalist agents. He has also co-created a new conference for reinforcement learning research.

Université de Montréal and Quebec AI Institute (Mila), CA

Aditya Bondada

Brookhaven National Laboratory, USA

Talk title: Toward Self-Driving Beamlines: Embodied Intelligence for Scientific Automation

Abstract: Synchrotron experiments are often limited by manual, repetitive workflows that constrain throughput, reproducibility, and experimental flexibility. We present a modular robotic platform that brings embodied intelligence to beamline operations at the National Synchrotron Light Source II (NSLS-II). Developed in collaboration with the Center for Functional Nanomaterials (CFN) and the Complex Materials Scattering (CMS) beamline at Brookhaven National Laboratory, the Extensible Robotic Beamline Scientist (ERoBS) combines robotic manipulation with AI-driven control to support autonomous, adaptive experimentation. The system integrates ROS2 with the Bluesky experimental orchestration framework, enabling real-time coordination of robotic tasks and experimental procedures. It supports a range of end-effectors -including vacuum and parallel-jaw grippers, and custom precision liquid dispensing tools - with an automatic tool exchange feature that facilitates seamless transitions between tools. Real-time environmental mapping and digital twin simulation allow for dynamic obstacle avoidance, intelligent task planning, and collision-free execution. These capabilities enable safe and responsive operation in complex, evolving beamline environments while supporting user-defined workflows tailored to specific sample types and constraints. By embedding intelligence in both physical action and decision-making, this platform enables in-situ, fully autonomous X-ray experiments, advancing the speed, precision, and accessibility of materials research at NSLS-II.

Bio: Aditya Bondada is a robotics researcher at Brookhaven National Laboratory, specializing in robotics and autonomous systems for experimental science. He holds a Master's in Robotics from Northeastern University, where he was part of the Silicon Synapse Lab, developing bio-inspired multimodal robots.

Page updated

Google Sites

Report abuse