SUM specialists will train a neural network to select materials for authors of scientific publications
A researcher is expected to be able to compile a specific ranking of useful materials that the system will factor in when adjusting the list of proposed articles at the time of repeated and subsequent requests.
The State University of Management is developing a digital platform that will help scientists automatically select the most appropriate materials for their studies, both from related and completely unrelated fields of study, TASS learned from the university’s press service.
The Digital Economy and High Technologies project laboratory is working on a platform solution that will make it possible to consolidate materials from libraries, scientific databases and journals into a single collection and, with neural network smart search, will help researchers quickly and easily select materials for their studies from different scientific fields, adjusting the principles of selection to their needs. The system allows users to download the entire text of an article for analysis, which is impossible in conventional search and database systems,
SUM’s press service explained to TASS.
A researcher is expected to be able to compile a specific ranking of useful materials for the system to factor in when adjusting the list of proposed articles (the backward error method) during repeated and subsequent requests. The developers believe that the primary purpose of the system is to advise research authors on nontrivial ways of its development. In particular, for a researcher developing new solutions in the field of construction, the system will offer articles in biology on weight distribution in animals.
Unexpected analogies with the animal world can find their application in actual scientific projects. In addition, the selection of materials from different fields contributes to the development of technologies across disciplines,
TASS learned from Mikhail Nachevskiy, the project author, Head of the Digital Economy and High Technologies project laboratory.
The authors are already testing the neural network algorithms developed using SUM’s electronic library and are negotiating cooperation with representatives of the Russian State Library and the CyberLeninka online scientific library. It is expected that the first users of the system will be university teachers. In the future, the university plans to bring the project to the inter-university level and create a specific social network where scientists would be able not only to find each other’s publications but also to agree on joint research.
We would certainly not like to limit ourselves to the Russian-speaking segment of scientific materials. The developed platform may by adapted at the later stages of project implementation for English-language searches,
said SUM’s representative.
Подпишись на тг-канал "Наш ГУУ"
Дата публикации: 12.11.2019