Scholarlink

Abstract <p>The work is devoted to the problem of adaptation of language models to a local mathematical subject area. On the example of ontology and knowledge graph for ordinary differential equations, queries in natural language are formed. The knowledge graph is created by means of intelligent data analysis in the semantic library of subject areas LibMeta. Adaptation is achieved by limiting a large language model when integrating with the knowledge graph of the subject area. The goal of the work is to create an environment for using a digital assistant in Russian when mastering scientific knowledge in a local subject area and scientific research. The limitations of the language model for the subject area of ordinary differential equations are realized by creating a set of templates and checking the truth of the answers based on them. Applications of the research results are expected to be implemented in systems of mathematical knowledge.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Navigation with Large Language Models in Subject Domain of Ordinary Differential Equation

  • 期刊论文

  • 发表日期：

    2025年6月1日

查看全文
  • O. M. Ataeva,
  • N. P. Tuchkova

摘要

Abstract

The work is devoted to the problem of adaptation of language models to a local mathematical subject area. On the example of ontology and knowledge graph for ordinary differential equations, queries in natural language are formed. The knowledge graph is created by means of intelligent data analysis in the semantic library of subject areas LibMeta. Adaptation is achieved by limiting a large language model when integrating with the knowledge graph of the subject area. The goal of the work is to create an environment for using a digital assistant in Russian when mastering scientific knowledge in a local subject area and scientific research. The limitations of the language model for the subject area of ordinary differential equations are realized by creating a set of templates and checking the truth of the answers based on them. Applications of the research results are expected to be implemented in systems of mathematical knowledge.