Details

Razložljivo zaznavanje mrežnih vdorov z velikimi jezikovnimi modeli
ID Malenšek, Martin (Author), ID Pejović, Veljko (Mentor) More about this mentor... This link opens in a new window, ID Grohar, Miha (Comentor)

.pdfPDF - Presentation file, Download (507,10 KB)
MD5: 597112FB54BC326E7138BE81A781C218

Abstract
Omrežni vdori predstavljajo pomemben sodobni varnostni izziv, saj lahko zloraba ranljivosti informacijskih sistemov vodi v nepooblaščen dostop do občutljivih podatkov, motnje v delovanju kritičnih storitev, finančno škodo ter zmanjšanje zaupnosti, celovitosti in razpoložljivosti informacijskih sistemov. Strojno učenje omogoča samodejno zaznavanje omrežnih vdorov, vendar imajo tovrstni sistemi običajno dve veliki omejitvi: omrežne podatke obdelujejo »daleč« od mesta zajema ter uporabniku vrnejo zgolj oznako razreda (npr.\ DDoS, PortScan) brez razlage, na podlagi česa je bila odločitev sprejeta. V tem delu obravnavamo razložljivost takšnih napovedi z velikimi jezikovnimi modeli (LLM) z izvedbo na pametni omrežni kartici (SNIC), oziroma neposredno v omrežju. V delu zasnujemo in implementiramo sistem FlowExplain, ki iz napovedi enostavnega klasifikatorja in značilk omrežnega toka samodejno, s pomočjo LLMa, ustvari razlage. FlowExplain temelji na prilagojenem (angl.\ \textit{fine-tuned}) manjšem lokalnem modelu, pri čemer se za prilagajanje uporabljajo posebej zbrani, s strani strokovnjakov označeni, učni primeri omrežnih vdorov. Z FlowExplainom pokažemo, da je avtomatizirano, razložljivo zaznavanje omrežnih vdorov na lokalni infrastrukturi izvedljivo v praksi.

Language:Slovenian
Keywords:Zaznavanje omrežnih vdorov, Razložljiva umetna inteligenca, Veliki jezikovni modeli, Klasifikator XGBoost, LoRA prilagajanje
Work type:Bachelor thesis/paper
Typology:2.11 - Undergraduate Thesis
Organization:FRI - Faculty of Computer and Information Science
Year:2026
PID:20.500.12556/RUL-189574 This link opens in a new window
COBISS.SI-ID:304413699 This link opens in a new window
Publication date in RUL:09.10.2026
Views:211
Downloads:51
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Explainable Network Intrusion Detection with Large Language Models
Abstract:
Network intrusions attacks represent a significant modern security challenge, as exploitation of vulnerabilities in information systems can lead to unauthorized access to sensitive data, disruption of critical services, financial damage, a reduction in the confidentiality, integrity, and availability of any information system. Machine learning enables automated detection of network intrusions; however, such systems typically suffer from two major limitations: network data is processed \enquote{far} from the point of capture, and the user is returned only a class label (e.g., DDoS, PortScan), without any explanation of the basis for the decision. In this work, we address the explainability of such predictions using large language models (LLMs) implemented on a smart network interface card (SNIC) directly in the network. We design and implement the FlowExplain system, which automatically generates explanations from the predictions of a simple classifier and network flow features with the help of an LLM. FlowExplain is based on a fine-tuned, smaller local model, for which specially collected training examples of network intrusions, labeled by experts, are used for fine-tuning. With FlowExplain, we show that automated, explainable network intrusion detection on local infrastructure is feasible in practice.

Keywords:Network intrusion detection, Explainable AI, Large language models, XGBoost classifier, LoRA fine-tuning

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back