Software for virtual tours

Abstract

The article discusses the methods and means of implementing a dialogue system using modern technologies of natural language processing, which will allow visitors to virtual art exhibitions to receive information about the exhibition, author and works in the form of an interactive conversation, which is built on the rules educational function in the "Exhibition" system, which modern web services of virtual exhibitions do not support. The architectures of dialog systems, text classifiers, corpora of Ukrainian texts, means of implementation of dialog systems for web applications in JavaScript and Python are analyzed. Based on their capabilities, disadvantages and advantages, the tools for developing a virtual guide were chosen. The existing systems of artificial intelligence for excursions, namely the works of "Promobot" and "Pepper", are considered, as well as the existing dialogue systems "Siri" and "Alice". The necessity of creating the module "Guide" in the system "Exhibition" is revealed and substantiated. The architecture of the dialog system for conducting virtual tours in cooperation with the web service of art exhibitions has been designed. The result is a dialogue system "Exi", which was integrated into a web service for organizing and conducting art exhibitions "Xolst". It has been experimentally proven that a virtual guide implements the necessary functionality for conducting tours, increases user interest in the exhibition, improves the perception of artwork and creates a positive emotional environment. Ref. 29, pic. 5.

Keywords

• dialog system
• dialog manager
• natural language processing
• virtual tour
• artistic work
• graphs of knowledge