Abstract High construction costs and unstable evaluation signals remain challenges in cross-task prompt engineering. Existing methods struggle with diversity maintenance at scale. To address these issues, this paper proposes EvoPrompt-ELO, an automated prompt-generation framework based on evolutionary strategies and ELO Rating System (ELO)-based dynamic adversarial evaluation. The framework employs Large Language Models (LLMs) as mutation operators to generate candidate prompts, incorporates Sentence-BERT (Sentence Bidirectional Encoder Representations from Transformers) for diversity constraints, and utilizes an LLM-as-a-Judge (treating an LLM as an evaluator) pairwise comparison mechanism combined with ELO scoring to achieve dynamic quality assessment and survival-of-the-fittest selection. Experiments show that the generated prompts achieve a 61.08 % win rate against the seed prompts in LLM-based blind testing, achieving broad semantic-mode coverage – doubling the number of distinct semantic clusters over the seed set (from 10 to 20) – while quality-guided selection concentrates the corpus on high-value regions, and successfully construct a task-agnostic, high-quality general-purpose prompt corpus Prompt-8K (8,261 prompts) efficiently. Furthermore, the ELO-based dynamic evaluation mechanism provides a trustworthy and scalable quality assessment framework that aligns with principles of adaptive trust management in evolving AI systems, enabling progressive refinement of knowledge repositories through principled adversarial evaluation and rating aggregation.
The study aimed to develop the concept of an innovative computer system with a dynamic architecture based on monocopters that function as distributed computing nodes. The research methodology included an analysis of design principles, modelling of the activation and energy distribution process, as well as the study of cooling capabilities and adaptive control of the system structure. As a result of the study, a concept of a computer system consisting of monocopters that perform the functions of computing nodes was developed. The key components of these monocopters were identified their structure, which includes local computing modules, power supplies and communication facilities for wireless communication. Activation of the system using external microwave power transfer, which became the basis for providing power to monocopters and initiating their functions, was emphasised. The study noted that the microwave energy entering the inductive coils can be converted into an electric current, providing the initial power to the system. The idea also envisages that monocopters can determine their position relative to other elements using a wireless data bus, allowing for an ordered three-dimensional structure that adapts to external conditions and requirements. The study indicates the potential of the system for the dynamic organisation of computational tasks that can be distributed among monocopters. This improves the reliability and efficiency of the system in the event of a change in topology or failure of individual nodes. Additionally, the study noted that the system can be scaled without delays in operations, creating new opportunities for the application of this technology in various fields, including robotics, adaptive materials and technologies with reduced dependence on conventional energy storage. Thus, the study results indicate the possibility of creating an adaptive and mobile computer system that can effectively respond to changing external conditions and provide high data processing performance.
Stress is a growing issue in modern workplaces, affecting mental well-being, productivity, and health. This paper explores the use of computer mouse movement patterns as non-invasive behavioral indicators of psychological stress. The study presents a system for detecting stress levels based on mouse interaction data. An experimental task was designed to induce different stress levels. The collected data were used to train machine learning models, several of which achieved high classification accuracy. The results demonstrate the potential of mouse-based behavioral analysis as a practical approach to real-time workplace stress detection. This research contributes to the development of accessible stress monitoring tools and highlights the value of behavioral data analysis for improving occupational health.
The task of authorship attribution is a classification problem where a given text is analyzed by a machine learning algorithm with the goal of verifying its author. This task has great application in the world of forensic analysis and intellectual property problems. In this work, we propose to use Support Vector Machines paired with Term Frequency - Inverse Term Frequency to solve authorship attribution in cases where the author is part of a predetermined set of authors, called a closed set. This method could also be applied to large language models in order to find out if those have a "style" of writing. In the end, we show how we compared our results with related works and how we managed to improve the results on a well known benchmark dataset by 6.4 %.
