Search academic texts

Information × Registration Number 0225U001774, (0123U103228) , R & D reports Title Develop methods and tools using a virtual dialog assistant for digital agriculture and environmental and medical projects popup.stage_title Розробити архітектуру і алгоритми роботи діалогових систем типу віртуального помічника, що використовують у своїй роботі графові бази знань онтологічного типу; залучення методів глибокого навчання для аналізу вихідних природно-мовних фраз користувача при автоматизованому створенні формальних запитів до онтології; адаптація діалогових систем до певних предметних областей, створення і тестування робочих прототипів. Head Anna A. Litvin, Registration Date 07-02-2025 Organization V.M. Glushkov Institute of Cybernetics of the National Academy of Sciences of Ukraine popup.description1 The purpose of the work is to develop a technology for creating a virtual dialog assistant for agriculture digitalization, environmental and medical projects: principles development of named entities automated retrieval and semantic connections between them through natural language texts analyzing involving tools based on neural networks. popup.description2 The aim of the research is to develop methods and tools for obtaining information in natural language with the help of a virtual dialog assistant, using data obtained from wireless sensor networks to express diagnostics of plant conditions, environmental monitoring, and human health. During the reporting period, the following results were obtained: 1) A multi-agent architecture of the dialog system was developed, which contains an ontology that allows distributing tasks among independent agents. 2) A comprehensive methodology for using an ontology-driven structured instruction system in interaction with ChatGPT was investigated. A method was formalized that involves meta-learning to create and configure structured prompts and context ontology, as well as a sequence of operations for intent detection, named entity identification, context selection, and hint instruction generation that allows ChatGPT to be provided with specific information by providing selected contexts, which can improve its knowledge in specific subject areas and improve its performance when working with data represented in an inflected language. 3) To provide medical rehabilitation support, an algorithm for creating an informational natural language dialogue system that does not include neural network models was considered. This approach has proven its high efficiency compared to large language models. 4) The adaptation of the knowledge base of the dialog assistant to work with data received from a wireless sensor network to express diagnostics of plant conditions, environmental monitoring, and human health was considered in the context of using the collected data as a source of information for the dialog assistant. Product Description popup.authors Kedych Anna V. Litvin Anna A. popup.nrat_date 2025-02-07 Close