A professor at Boston University quantified ARES versus traditional experiment campaign design in his own research and realized a 60-times increase in speed. The tool is still in its early days Maruyama emphasizes. ARES OS is being evaluated by ten partners including Carnegie-Mellon University, the University of Missouri, and several commercial groups. “It was really the ability to use artificial intelligence and autonomy algorithms to design the next experiment, to target the next research goal.”ĪFRL currently has four different fully autonomous ARES robots running for various programs. The benefits of speeding experimentation by 100-times are obvious but that wasn’t the most important gain Maruyama says. Once he and other AFRL researchers began employing ARES, they went from doing roughly one physical/hardware experiment per day to 100 experiments. Researchers can write their own software modules to link their specific lab hardware (airflow sensors, chemical sensors, lasers, catalysts, micrometers, rate-controllers, even virtual simulators) via ARES OS’ plug-ins for their own research experiments.Īfter this link is established, the software can design experiments autonomously and perform experiments in a rapid, precise manner, learning how to achieve the research goal using AI. “Much of the challenge was building the software to talk to the hardware,” Maruyama says.
That’s where ARES differs.ĪRES OS features a core or a universal module with plug-ins that connect to experimental tools or scientific instruments such as electron microscopes. Such code had limited adaptability to new experiments. In the past, researchers wrote similar automation code just to get one job done, a one-time experimental process Maruyama explains. It’s essentially what ARES does, providing a basis to design experiments, answer research questions, execute experiments and analyze the results, which then helps design a new round of experiments. You say, ‘something happened there so let me go close to that ship and try to find something else there.’ You choose your next shot based on your previous results.” “You might start off searching for a ship but once you hit something, you don’t keep randomly. “Wouldn’t it be great if you could do experiments more like you were playing Battleship?” Maruyama thought. He and his colleagues quickly recognized that developing an autonomous tool could also yield insights to shape, and self-direct, follow-on experiments.
Maruyama sought initially to apply automation to speed his own materials research, growing carbon nanotubes in the lab. Using the artificial intelligence and autonomy algorithms built into ARES OS, researchers can transform experiments into autonomous “research robots” capable of directing and conducting their own research.ĭr.
It is available for download - free of charge - from AFRL. The result is the Autonomous Research System, an open source software operating system (ARES OS).
0 kommentar(er)
