Anisa Rula
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Anisa Rula has a total of 64 co-authors. Here are the top co-authors by number of shared publications:
- Devis Bianchini
University of Brescia
8 shared publications - Matteo Palmonari
University of Milano-Bicocca
8 shared publications Andrea Maurino
7 shared publications- Ada Bagozi
University of Brescia
6 shared publications - Blerina Spahiu
University of Milano-Bicocca
6 shared publications - Maria Angela Pellegrino
University of Salerno
5 shared publications
Latest Publications
- Open Access
- 6 July 2026
- Published by Elsevier in SSRN Electronic Journal
KGtelescope: Knowledge Graph Guided Model Routing for Factual Question Answering
Knowledge graphs (KGs) can serve generative AI not only as retrieved evidence, but also as an explicit decision layer for allocating model capacity. This paper asks whether KG-derived structural descriptors can guide routing from smaller to larger LLMs for factual question answering. We introduce KGtelescope, which constructs questions from typed KG relations and associates each with a structural profile covering relation pattern, query direction, support, entity prominence, and answer cardinality. Six Gemma and Llama models are evaluated on 131{,}616 DBpedia-derived questions under zero-shot and few-shot prompting, producing 1{,}579{,}392 responses. Factual accuracy varies strongly across KG-derived conditions. Query direction and answer cardinality have the largest effects; entity prominence (PageRank quartile) consistently predicts accuracy, with Q1 entities yielding up to 5.97 pp higher accuracy than Q3 across all 12 model configurations. For single-answer questions, the KG-informed routing policy significantly outperforms matched-budget random allocation ($\Delta = 5.26$ pp, $p < 0.001$), capturing 57.1\% of the always-large accuracy gain at a 25\% large-model budget. The routing signal transfers across model families, though transfer to unseen KG pattern families is weaker. KGtelescope positions KG structure as an interpretable, resource-aware routing layer for generative AI systems.
- 6 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessKGtelescope: Knowledge Graph Guided Model Routing for Factual Question Answering
Knowledge graphs (KGs) can serve generative AI not only as retrieved evidence, but also as an explicit decision layer for allocating model capacity. This paper asks whether KG-derived structural descriptors can guide routing from smaller to larger LLMs for factual question answering. We introduce KGtelescope, which constructs questions from typed KG relations and associates each with a structural profile covering relation pattern, query direction, support, entity prominence, and answer cardinality. Six Gemma and Llama models are evaluated on 131{,}616 DBpedia-derived questions under zero-shot and few-shot prompting, producing 1{,}579{,}392 responses. Factual accuracy varies strongly across KG-derived conditions. Query direction and answer cardinality have the largest effects; entity prominence (PageRank quartile) consistently predicts accuracy, with Q1 entities yielding up to 5.97 pp higher accuracy than Q3 across all 12 model configurations. For single-answer questions, the KG-informed routing policy significantly outperforms matched-budget random allocation ($\Delta = 5.26$ pp, $p < 0.001$), capturing 57.1\% of the always-large accuracy gain at a 25\% large-model budget. The routing signal transfers across model families, though transfer to unseen KG pattern families is weaker. KGtelescope positions KG structure as an interpretable, resource-aware routing layer for generative AI systems.
- 7 June 2026
- Published by Springer Nature
- p. 176-184
D3SD: A Web-based Tool for Data Service Design in Cyber-Physical Production Networks
- 7 June 2026
- Published by Springer Nature
- p. 176-184
D3SD: A Web-based Tool for Data Service Design in Cyber-Physical Production Networks
- 7 June 2026
- Published by Springer Nature
- p. 63-71
OntoLLM: A Framework for Trustworthy LLM Pipelines
- Top Cited PapersOpen Access
- 23 April 2026
- Published by Wiley in Advanced Intelligent Systems
OntoLogX: Ontology‐Guided Knowledge Graph Extraction From Cybersecurity Logs With Large Language Models
System logs represent a valuable source of cyber threat intelligence (CTI), capturing attacker behaviors, exploited vulnerabilities, and traces of malicious activity. Yet their utility is often limited by a lack of structure, semantic inconsistency, and fragmentation across devices and sessions. Extracting actionable CTI from logs, therefore, requires approaches that can reconcile noisy, heterogeneous data into coherent and interoperable representations. We introduce OntoLogX, an autonomous AI agent that leverages large language models (LLMs) to transform raw logs into ontology‐grounded knowledge graphs (KGs). OntoLogX integrates a lightweight log ontology with retrieval augmented generation and iterative correction steps, ensuring that generated KGs are syntactically and semantically valid. Beyond event‐level analysis, the system aggregates KGs into sessions and employs a LLM to predict MITRE ATT&CK tactics, linking low‐level log evidence to higher‐level adversarial objectives. We evaluate OntoLogX on both public and real‐world honeypot datasets, demonstrating robust KG generation across multiple LLMs backends and accurate mapping of adversarial activity to MITRE ATT&CK tactics. Results highlight the effectiveness of the methodology in constructing ontology‐compliant KGs, along with their value in extracting actionable CTI.
- 23 April 2026
- Published by Wiley in Advanced Intelligent Systems
Top Cited PapersOpen AccessOntoLogX: Ontology‐Guided Knowledge Graph Extraction From Cybersecurity Logs With Large Language Models
System logs represent a valuable source of cyber threat intelligence (CTI), capturing attacker behaviors, exploited vulnerabilities, and traces of malicious activity. Yet their utility is often limited by a lack of structure, semantic inconsistency, and fragmentation across devices and sessions. Extracting actionable CTI from logs, therefore, requires approaches that can reconcile noisy, heterogeneous data into coherent and interoperable representations. We introduce OntoLogX, an autonomous AI agent that leverages large language models (LLMs) to transform raw logs into ontology‐grounded knowledge graphs (KGs). OntoLogX integrates a lightweight log ontology with retrieval augmented generation and iterative correction steps, ensuring that generated KGs are syntactically and semantically valid. Beyond event‐level analysis, the system aggregates KGs into sessions and employs a LLM to predict MITRE ATT&CK tactics, linking low‐level log evidence to higher‐level adversarial objectives. We evaluate OntoLogX on both public and real‐world honeypot datasets, demonstrating robust KG generation across multiple LLMs backends and accurate mapping of adversarial activity to MITRE ATT&CK tactics. Results highlight the effectiveness of the methodology in constructing ontology‐compliant KGs, along with their value in extracting actionable CTI.
- Open Access
- 25 March 2026
- Published by Association for Computing Machinery (ACM) in Journal of Data and Information Quality
- Vol. 18 (1) , 1-40
Lost in LOD: Analyzing the Linked Open Data Cloud Quality Maze
Since 2007, the Linked Open Data (LOD) Cloud has served as a central hub for datasets following Linked Data (LD) principles, offering a large repository of interconnected information. Over time, it has undergone multiple quality assessments to ensure datasets are accessible, well-maintained, and meet standards. Grounded on metadata assessment performed over time, this paper examines the current quality of the LOD Cloud by analyzing 1,658 datasets from the December 2024 snapshot, evaluated against 52 quality metrics. By proposing a reproducible methodology, it reports about the quality assessment and the trend analysis assessing progress, identifying persistent problems, and verifying how datasets registered in the LOD Cloud evolve over time. According to results, many earlier issues persist. Datasets still lack consistency in metadata structure, licenses, and distribution format. Moreover, they mainly remain in archived versions, with real-time access often poorly maintained. Well-curated, up-to-date datasets are exceptions rather than the rule.
- 25 March 2026
- Published by Association for Computing Machinery (ACM) in Journal of Data and Information Quality
- Vol. 18 (1) , 1-40
Open AccessLost in LOD: Analyzing the Linked Open Data Cloud Quality Maze
Since 2007, the Linked Open Data (LOD) Cloud has served as a central hub for datasets following Linked Data (LD) principles, offering a large repository of interconnected information. Over time, it has undergone multiple quality assessments to ensure datasets are accessible, well-maintained, and meet standards. Grounded on metadata assessment performed over time, this paper examines the current quality of the LOD Cloud by analyzing 1,658 datasets from the December 2024 snapshot, evaluated against 52 quality metrics. By proposing a reproducible methodology, it reports about the quality assessment and the trend analysis assessing progress, identifying persistent problems, and verifying how datasets registered in the LOD Cloud evolve over time. According to results, many earlier issues persist. Datasets still lack consistency in metadata structure, licenses, and distribution format. Moreover, they mainly remain in archived versions, with real-time access often poorly maintained. Well-curated, up-to-date datasets are exceptions rather than the rule.
- Open Access
- 31 December 2025
- Published by Elsevier in Journal of Industrial Information Integration
- Vol. 50
A multi-layered data service model for Cyber-Physical Production Networks
- 31 December 2025
- Published by Elsevier in Journal of Industrial Information Integration
- Vol. 50
Open AccessA multi-layered data service model for Cyber-Physical Production Networks
- 29 October 2025
- Published by Springer Nature
- p. 23-41
Are Quality Dimensions Correlated? An Empirical Investigation Over Linked Data
- 29 October 2025
- Published by Springer Nature
- p. 23-41
Are Quality Dimensions Correlated? An Empirical Investigation Over Linked Data
- Open Access
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2112-2115
The Web of Data Quality Workshop
- 8 May 2025
- Published by Association for Computing Machinery (ACM)
- p. 2112-2115
Open AccessThe Web of Data Quality Workshop
- 28 January 2025
- Published by Springer Nature
- p. 276-280
KGHeartBeat: A Knowledge Graph Quality Assessment Tool
- 28 January 2025
- Published by Springer Nature
- p. 276-280
KGHeartBeat: A Knowledge Graph Quality Assessment Tool