Information security compliance has become critical for organizations worldwide, with the ISO/IEC 27000 family serving as the most widely adopted framework for establishing information security management systems. Despite their global acceptance, these standards present significant interpretation challenges due to their formal language, abstract structure, and extensive cross-referencing across 97 documents. Traditional retrieval-augmented generation (RAG) systems, which rely on independent text chunking and dense vector retrieval, prove inadequate for such highly interconnected regulatory materials, often fragmenting contextual relationships and reducing accuracy. This study introduces a privacy-preserving RAG framework that integrates LightRAG, a knowledge graph-based retrieval system, with locally hosted open-source language models. Unlike chunk-based RAG systems that treat document segments independently, the system in this study constructs a semantic knowledge graph that explicitly models relationships between clauses through typed edges representing cross-references, semantic similarity, and hierarchical dependencies. To enable rigorous evaluation, we developed a curated benchmark dataset of 222 multiple-choice questions with authoritative ground-truth answers, systematically constructed from official ISO standards, certification preparation materials, and academic sources. Through systematic evaluation on this benchmark, we show that knowledge graph-based retrieval achieves higher accuracy than chunk-based RAG and non-retrieval LLM baselines within the evaluated setup. The analysis indicates that embedding model quality is strongly associated with system performance, that hybrid retrieval modes combining local and global graph traversal tend to yield better accuracy, and that mid-sized open-source models paired with strong retrievers can approach the performance of larger proprietary systems. The best configuration achieves 90.54% accuracy, demonstrating the promising effectiveness of graph-structured retrieval for multiple-choice regulatory questions.

Understanding how people perceive and value landscapes is essential for sustainable planning and conservation; yet, traditional methods remain limited in scale and scope. This study introduces artificial intelligence (AI)-Perceptual Landscape Mapping (AI-PLM), an integrated analytical framework that combines geospatial intelligence, machine learning, and natural-language processing (NLP) to model collective human perception from social-media data. Using nearly 29 000 geotagged Flickr photographs and 148 000 user comments from Romania, AI-PLM operationalizes perception through three components: (1) Data collection and processing (systematic collection and normalization of multilingual, multimodal content), (2) AI-Spatial Cognition (identification of perception hotspots via Head/Tail Breaks and DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering combined with viewshed analysis), and (3) Affective–Semantic Intelligence (sentiment and topic modeling using transformer-based NLP). Results reveal strong spatial hierarchies of landscape appreciation, with intensity peaks in the Carpathians, Braşov, Bucharest, Maramureş, and the Black Sea coast. Sentiment analysis shows predominantly positive emotions associated with nature-oriented regions, while topic modeling highlights the prevalence of themes related to photography, heritage, and recreation. Together, these multimodal insights demonstrate a clear relationship between visibility, spatial clustering, and affective tone. The AI-PLM framework, thus, bridges physical geography and emotional expression, providing a scalable and transferable methodology for assessing cultural ecosystem services. By translating unstructured digital traces into structured spatial and semantic indicators, it advances the understanding of human–landscape interactions and offers practical tools for data-driven landscape management, conservation, and tourism planning in Romania and beyond.

Macedonian is a low-resource language for automatic speech recognition: annotated speech data are scarce, dialectal variation is substantial, and existing evaluations focus almost entirely on read speech in Standard Macedonian. This paper presents a unified pipeline that converts Macedonian speech, including regional dialects, into accurate transcripts and structured, queryable knowledge, a capability required for applications such as searchable parliamentary archives, broadcast transcription and subtitling, and dialectological documentation. Parameter-efficient adaptation of Whisper large-v3-turbo via low-rank adaptation is evaluated against strong zero-shot and language-specific baselines on four newly curated dialect corpora (Ohrid, Veles, Tikvesh, and Gostivar) and three Standard Macedonian corpora, two of which were collected for this work. The adapted model reduces word error rate by 57–70% relative to the strongest zero-shot baseline and by 33–66% relative to the language-specific BUKI Whisper 2.0 model on dialectal speech, with comparable improvements over zero-shot baselines on standard Macedonian speech, while updating only about 0.7% of parameters. Beyond transcription, the pipeline adds speaker diarization with cross-recording speaker linking and a graph-based retrieval-augmented generation component that enables speaker-, topic-, and time-aware querying of diarized transcripts, evaluated on long-form Macedonian parliamentary recordings. Together, these results establish parameter-efficient adaptation, speaker-aware processing, and graph-based retrieval as a practical and transferable framework for transforming under-resourced speech into accessible, structured knowledge.

Understanding how societies perceive and adapt to environmental change requires analyzing awareness as a dynamic process emerging from coupled human–environment interactions. This study presents a longitudinal, data-driven analysis of environmental awareness based on three years of social-media discourse related to air pollution. Using sentiment analysis (Valence Aware Dictionary and Sentiment Reasoner and Twitter-RoBERTa) and topic modeling (BERTopic), we quantify the temporal evolution of collective public awareness and identify dominant discourse themes. A correlation analysis between sentiment indicators and PM10 levels reveals synchronized fluctuations, while mutual information indicates that public awareness becomes increasingly dependent on or shaped by pollution variability. Furthermore, Kruskal–Wallis tests confirm statistically significant temporal variations in sentiment distributions, underscoring adaptive shifts in public awareness across years. By interpreting awareness as a measurable order parameter within a complex socio-environmental system, this work demonstrates how collective perception emerges from nonlinear feedbacks between environmental forcing and public information flow. The proposed multilayer socio-environmental perception framework integrates machine-learning-based sentiment and topic analysis with complex-systems concepts to quantify emergent awareness dynamics from longitudinal social-media data.