Library Information Search in the Age of Digital Transformation: The Potential of Neural Networks and Generative Artificial Intelligence

Abstract

Today, as the information society evolves and libraries undergo digital transformation, there is a pressing need to find a balance between the vast volume of electronic information and the personalized needs of users. Existing traditional information retrieval models are no longer able to cope with the complex networks of user interactions with information resource objects and do not fully explore the deep connections between user preferences and content. The purpose of this study is to review modern technologies for implementing information retrieval in libraries using neural networks and artificial intelligence. The analysis of international models – graph neural networks, deep neural networks based on embedding, neural collaborative filtering, singular value decomposition, relational graph hyperfine networks, hybrid K-nearest neighbors, deep neural networks, and hyperfine neural networks – has identified key implementation criteria for these models, namely: architecture specifics, accuracy metrics, performance evaluation, implementation methodology, library type and type, conceptual authors, and other criteria. As a result of the conducted analysis, it is proposed to create a model that will combine traditional technologies of automated library and information systems with integration into a complex neural network model with algorithms that allow taking into account the personalized needs of users.