New Protocol System Development to Scale Scientific Robotic Agents

The aim of automate scientific research is to approach the reality by developing artificial intelligence agents that perform tasks in physical environments. However, the lack of reliable assessment environments to test these agents is limited to the progress in the field. The researchers developed a new system to solve this problem.

• SciHorizon-eLab was a new platform that automatically converts scientific protocols to laboratory tasks Tags
• The research team aims to create scalable assessment environments, significantly reducing the time and workload of the manual task design
• The system can be understood directly by the machines of scientific protocols, unlike the simulation-based laboratory tests that exist, and translate into the forms that can be realized
• The platform allows for systematic testing of a wide range of scientific experiments using physical robots or simulated agents

This development addresses one of the most important problems of automation work in scientific research. Previously, it takes a lot of time to program laboratory tasks manually and blocked scalable solutions. The protocol compiler of SciHorizon-eLab allows researchers to test agents on a number of and diverse tasks.

The system can be a milestone to evaluate the capabilities of embolised agents more comprehensively, and can open the door of new works in the scientific automation field.