Knowledge transfer methodology: focus on neural networks
Abstract
The article analyzes the mechanisms of knowledge transfer in the context of dynamic development of neural network technologies. The focus is on the methodological principles of knowledge transfer, the theoretical construct of which is the relay model of M.A. Rozov. Knowledge, considered as a social program formed and improved in the process of joint community activities, discussions and exchange of experience, is not limited to a simple set of facts and data, but includes norms, values, research methods, traditions as an integral part of its complex structure. Neural networks, being new and fundamentally different participants in knowledge relays, are able to effectively process information, remaining limited by the framework of formal representation. Machine «knowledge» is deprived of intentionality necessary for a deep understanding of the context, which gives rise to the risk of reducing meanings to a superficial set of patterns. Particular attention is paid to the problem of hidden standardization, when neural networks reproduce established discursive models, depriving the intellectual space of the opportunity for critical reassessment. For a modern university, faced with a revision of the content of existing educational practices, given the expansion of neural networks, the topic of the presence of a philosophical filter, shown through the prism of the course «Logic and Critical Thinking», is becoming more relevant. In conclusion, it is noted that the widespread use of neural networks insistently requires the disclosure of issues of knowledge production within the framework of the updated digital environment.