The development of the climate observations question-answer (QA) information system relies on heterogeneous climate data in various formats (text, numerical, graphic, video, audio, geographic and monitoring data). A mandatory element of such a system is a tool that allows processing and analyzing such data.
Searching and retrieving data is a central part of the system in question, since the quality of the generated answer heavily depends on it. The exact way the data is retrieved is critical to the output of a QA system as well as to decision-making problems, since there are situations in which the LLM generates a contextually appropriate but factually incorrect answers that do not match the input. Using correct metrics and algorithms for some data types and incorrect ones for others can cause the permissible threshold of irrelevant data to be exceeded, which in turn can cause the quality of the answers to decrease. Retrieval-augmented generation (RAG) systems can also be used to optimize input data for that task.
This work discusses various algorithms for data extraction and document ranking, as well as the possibility of using ensembles of LLM agents in development of the QA system that works with climate data.

The article is devoted to the issues of digitalization of the lecture form of work at a university. It is shown that a number of problems associated with modern social and personal transformations lead to a decrease in the effectiveness of the classical lecture. A new approach to the lecture process is proposed, based on the use of multimedia lectures with an intelligent pedagogical agent in the educational process. It has been shown that the use of such lectures can help improve the effectiveness of training sessions. To practically test the effect of applying the approach, a training system is designed and implemented to ensure the creation and use of multimedia lectures. Information is provided about the information model underlying the system and its cloud architecture. The operation of a demonstration prototype of a multimedia lecture player, including the functionality of a pedagogical agent, is described. It is concluded that it is possible to build a player, including the functionality of a pedagogical agent running on consumer-level computer devices. Plans for further research are presented.

The paper discusses the use of neural network technology to support decision-making in the educational process of a university using a cognitive learning model. A software solution has been developed for a digital profile of a student based on an electronic portfolio of students, using artificial intelligence algorithms, modern web technologies, as well as cognitive learning models. The neural network was trained on prepared student data, which was obtained using a specially developed psychodiagnostic complex. Using a digital profile allows students to track their learning process based on recommendations offered by a neural network, make optimal decisions, build personalized educational trajectories, and also adjust educational learning trajectories.

The article is devoted to the analysis of the role and efficiency of neural network algorithms in the tasks of automatic abstracting and summarization of texts, which are key in the field of natural language processing (NLP). The main goal of automatic abstracting is to extract and generate essential information from texts to provide quick access to the main content without having to read the whole document. The paper discusses the main challenges faced by developers in implementing abstracting algorithms, including understanding context, irony, maintaining text cohesion, and adapting to different languages and styles. Special attention is given to neural network models such as Transformer, BERT, and GPT, which have shown outstanding performance in automatic text abstracting due to their ability to learn on large amounts of data. The article also highlights the contributions of leading researchers in the field of deep learning and analyzes the methods underlying state-of-the-art NLP algorithms, highlighting the importance of continuous technological progress in improving abstracting quality and information accessibility. The article will be of interest to a wide range of readers, including researchers in the field of artificial intelligence and NLP, software developers engaged in automation of text processing, as well as specialists in areas where fast processing and analysis of large amounts of textual information is required, such as legal practice, medical diagnostics and scientific research. In addition, the material of the article will be useful for teachers and students studying data processing and artificial intelligence technologies, providing them with actual examples of applying theoretical knowledge in practical projects.

This article discusses the selection of the optimal neural network model YOLOv8 (YOLOv8s, YOLOv8l, YOLOv8x, YOLOv8m, YOLOv8n) using the hierarchy process analysis method, which allows structuring and systematizing complex decisions based on multi-criteria assessments. The main focus is on identifying and comparative analysis of the most significant criteria for assessing the effectiveness of neural network models, such as training time, as well as the Precision, Recall and F1-score metrics. These metrics play a key role in computer vision tasks, especially when it comes to object detection. During the study, pairwise comparison matrices were constructed, which allow not only to visually represent the relative importance of each of the selected parameters, but also to quantitatively assess their impact on the overall effectiveness of the model. The process of forming pairwise comparison matrices includes the opinion of experts in the field of machine learning and computer vision, which ensures a high degree of reliability of the results. After processing the data and performing calculations, including weighting each criterion, priorities for alternative YOLOv8 models were derived. As a result of the calculations, it was revealed that the YOLOv8n neural network model has the highest priority among all the alternatives evaluated. This emphasizes its superiority compared to other modifications of this model.