Answers are grounded in this paper and cite their sources.
The cognitive capacity of modern manufacturing systems is, paradoxically, limited by an excessive focus on the use of machine- and deep-learning technologies for learning about processes and systems. These are learning models that are, in fact, primarily based on specific statistical processing of a large amount of data collected through the sensor system. However, two crucial features that are missing for achieving real cognitive capacity are the ability to acquire deep knowledge about the process that humans gain efficiently through experience, and to assign semantic relevance to the collected data for the given context of a process, which is carried out in a system. Mixed reality is potentially one of the tools that can be effectively used to capture human knowledge about a process and create associations between the collected data and the meaning these data have in the context of a process taking place in a system. The paper shows a case of using mixed reality in the form of a perceptual interface through which a human operator teaches a computer system about the assembling process of medium voltage switching equipment that it monitors. That learning process is carried out by building associations between the concepts of a semantic network that stores explicit, but also tacit knowledge about the process. Association formation is effectively done by mixed reality being applied by a human in the process. Finally, the cognitive capacity of the computer system to learn about the ongoing process autonomously and permanently is provided by applying inference-based analogies of the topological features of the semantic network associations.
No takes yet. Share an insight, caveat, or question.
Miloš et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: