Fig. 2: An illustration of a distributed self-driving laboratories (SDLs) architecture.

a Conceptual framework of components used to build a network of distributed SDLs for closed-loop optimisation. The framework encompasses a holistic integration of data, software, hardware, and workflow, taking into account the flow of information within cyberspace and materials within physical space. Initiated by the scientist specifications, these flows autonomously evolve across cyber and physical spaces until they accomplish the research goals or exhaust allocated resources. The illustration of the design of experiments was created by ref. ⁸¹. b Dynamic knowledge graph approach that is structured into three layers. The first layer represents the real world, where hardware is located and reactions take place. The second layer consists of a dynamic knowledge graph in cyberspace, hosting information such as the digital twin of the hardware and chemical data. The third layer comprises active agents that continually monitor the status of the knowledge graph, dynamically restructuring it, and actuating changes in the real world. The illustration of docker was created using flickr.com.