Implementation of Artificial Intelligence in Internet Search

The article presents an evaluation of the application of artificial intelligence in searching and retrieving information on the Internet. Potential sources of information were identified, procedures for searching Internet resources were developed, and methodologies for analyzing and presenting results were proposed. Natural Language Processing (NLP) methods applicable to identifying the semantics of texts, i.e. understanding the meaning and context contained in sentences and documents, were applied. Webcrawlers and a Python library were used to prepare the data. The Bert-Base-Uncased and Bert-Large-Uncased models were analyzed in detail, determining the generation times and probabilities of the correctness of answers, as well as the relevance parameter, as evaluations of whether the generated answer agrees with what the user wanted to search for.