ABSTRACT
Data structuring has grown increasingly complex in the age of information technology,1-6 when virtually all fields of science are recognized as interconnected. On the side of human-based information retrieval, the inflation of information available via electronic systems resulted in enormous quantitative load of relatively routine content-filtering tasks for human to perform, a price to pay for the efficiency of content access and delivery. Search engines have been developed in response to this problem; being universal and common are now the main keywords, as the algorithms developed for the most successful public systems include learning layer into search and contents provision. While such inbuilt evolutionary elements are crucial for public or commercial success of the search engines, in fact featuring soft functional separation, these generally degrade search precision in highly specific tasks, such as retrieval of scientific articles on a narrow subject, in case of which common methods prove extremely inefficient (delivery failure for tasks where user input is less valued than system output).
