Fig. 5: Autonomous workflow triggered in response to goal requests from scientists as information travels within the knowledge graph.
From: A dynamic knowledge graph approach to distributed self-driving laboratories
a The Reaction Optimisation Goal (ROG) Agent translates the specification of a scientist into a machine-readable statement and instantiates it into the knowledge graph. b The Reaction Optimisation Goal Iteration (ROGI) Agent initiates the design-make-test-analyse cycle, during which other agents query information from the digital twin and actuate the hardware. c The progress of goal pursuit is assessed after each iteration, determining whether to proceed to the next cycle. Steps (b) and (c) are iterated until either goals are achieved or resources are depleted.
