Repozitorij FER-a

Prilog u časopisu Otvoreni pristup

SHAP-GSD: Temporal Multi-Granular Explanation Method for Graph Neural Networks in Network Intrusion Detection

Luša, Riko Pintar, Damir Vranić, Mihaela

2026

In graph neural network (GNN)-based network intrusion detection systems (NIDS), an alert may reflect flow attributes, recent endpoint communications, or anomalous endpoint behaviour. Feature-only Shapley methods identify which flow attributes trigger…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva

IEEE open journal of the Communications Society Vol. 7, str. 11714–11739

DOI10.1109/ojcoms.2026.3734159 eISSN2644-125X

Repozitorij FER-a

Network intrusion detection system (NIDS) graph neural network (GNN) explainability Shapley values temporal graph learning

Prikaži detalje Otvori

Prilog u časopisu Otvoreni pristup

Improving Signal Quality in Non-Contact Electrocardiography: Novel Strategy for Motion Artifact Reduction

Stanešić, Antonio Klaić, Luka Cindrić, Dino Cifrek, Mario

2026-06-07

Capacitive electrocardiography (cECG) enables non-contact heart rate monitoring through clothing, but motion artifacts remain a critical limitation for practical applications. We present a novel motion artifact removal method using non-contact floati…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva

Sensors Vol. 26, Br. 12, str. 3643–3643

DOI10.3390/s26123643 eISSN1424-8220 pISSN1424-8220

Repozitorij FER-a

capacitive electrodes noncontact electrodes biopotentials ECG cECG biomedical signals gamification NLMS multi-referenced NLMS QRS PLI

Prikaži detalje Otvori

Prilog u časopisu Otvoreni pristup

Assessing Graph Data Serialization and Context Selection for Knowledge-Augmented Question Answering with Large Language Models

Horvat, Marko Kupres, Sven Leo Delimar, Marko

2026-10-07

Large language models (LLMs) can produce inaccurate answers even when relevant facts are supplied in the prompt. This study examines how graph data serialization, entity identifiers, and context selection affect factual question answering in knowledg…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za primijenjeno računarstvo

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za visoki napon i energetiku

Applied sciences (Basel) Vol. 16, Br. 19

DOI10.3390/app16199919 eISSN2076-3417

Repozitorij FER-a

large language models knowledge graphs ontologies knowledge representation data augmentation natural language processing

Prikaži detalje Otvori

Prilog u časopisu Pristup onemogućen

Robust evaluation of lexical, transformer, and retrieval-grounded language models for health condition identification

Bagić Babac, Marina

2026-09-14

Health condition identification from short symptom narratives remains a demanding information-systems task because useful solutions must combine representation quality, transfer robustness, retrieval support, and operational efficiency. This paper co…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva

Journal of intelligent information systems Vol. -, Br. -, str. -–-

DOI10.1007/s10844-026-01092-1 eISSN1573-7675 pISSN0925-9902

Repozitorij FER-a

Disease Models Information Model Language Processing Natural Language Processing (NLP) Machine Learning Computational Linguistics Natural Language Processing in Electronic Health Records

Prikaži detalje

Doktorski i znanstveni magistarski radovi Otvoreni pristup

Razvoj raspodijeljene programske potpore uvjetovan sigurnosno ograničenim tokom podataka"

Lovrenčić, Rudolf

2026-03-31

Sigurnost se često ističe kao glavni razlog zbog kojeg se organizacije suzdržavaju od korištenja računarstva u oblaku. Prilikom korištenja računarstva u oblaku za pohranu i obradu podataka, privatnost podataka može biti ugrožena na dva načina: i) u p…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za elektroniku, mikroelektroniku, računalne i inteligentne sustave

Repozitorij FER-a

multi-cloud security-constrained data flow data and code fragmentation heuristic optimization secure distributed computing protocols data obfuscation računarstvo u oblaku korištenjem više pružatelja usluga sigurnosno-ograničeni tok podataka fragmentacija podataka i programske logike heuristička optimizacija protokoli raspodijeljenog računarstva za sigurno računanje maskiranje podataka

Prikaži detalje Otvori

Prilog u časopisu Otvoreni pristup

MC-PanDA++: Simpler, Stronger, and More Robust Domain-Adaptive Panoptic Segmentation

Martinović, Ivan Šarić, Josip Asano, Yuki M. Šegvić, Siniša

2026

Unsupervised domain adaptation (UDA) reduces the annotation burden in panoptic segmentation by leveraging a cost-effectively labeled source domain (e.g., synthetic) and an unlabeled target domain to bridge the distribution gap. Existing panoptic UDA …

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva

International journal of computer vision Vol. 134, Br. 10

DOI10.1007/s11263-026-03046-0 eISSN1573-1405 pISSN0920-5691

Repozitorij FER-a

Panoptic segmentation · Domain adaptation · Confidence estimation · Domain shift

Prikaži detalje Otvori

Prilog u časopisu Otvoreni pristup

Test Methods for Validating Inductive Transformer Performance during Cable Discharge: Main Principles, Parameters and Testing Procedures

Grđan, Andrea Žiger, Igor Hanić, Zlatko Tannemaat, Imre Huth, Ralf Novko, Ivan

2024

High voltage connections will discharge through inductive voltage and combined transformers when they are switched off. In the past, there was a wide margin between the required capabilities and the actual capabilities of these components. Nowadays, …

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za elektrostrojarstvo i automatizaciju

Energies Vol. 17, Br. 4, str. 1–14

DOI10.3390/en17040799 pISSN1996-1073

Repozitorij FER-a

Keywords: inductive voltage transformers combined transformers line discharge test methods

Prikaži detalje Otvori

Prilog u časopisu Dostupan nakon (2028-10-01)

Explaining deceptive language in topical discourse: Integrating transformer models and explainable AI using LIME and SHAP for transparent text-based deception detection

Vrljić, Bjanka Botički, Ivica

2027

Detecting deception in textual communication is an important challenge in online environments where misleading information and deceptive narratives are increasingly prevalent. This study investigates the integration of fine-tuned transformer models, …

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za primijenjeno računarstvo

Information processing & management Vol. 64, Br. 2

DOI10.1016/j.ipm.2026.105213 eISSN1873-5371 pISSN0306-4573

Repozitorij FER-a

Natural language processing Transformers Large language models Explainable AI Deception detection

Prikaži detalje

Prilog u časopisu Otvoreni pristup

Adaptive Decision-Making Policies for Stochastic Electric Vehicle Routing with Time Windows Using Genetic Programming

Đurđević, Alan Frid, Nikolina Đurasević, Marko

2026-10-12

Intelligent transportation systems increasingly rely on electric vehicle fleets operating under uncertain conditions, where routing decisions must adapt to changes in customer demand, service duration, and travel time while satisfying battery, chargi…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za elektroniku, mikroelektroniku, računalne i inteligentne sustave

Sensors Vol. 26, Br. 20, str. 6457–6498

DOI10.3390/s26206457 eISSN1424-8220

Repozitorij FER-a

electric vehicle routing problem sensor-enabled fleet management stochastic vehicle routing online routing policies decision planning genetic programming hyper-heuristics

Prikaži detalje Otvori

Doktorski i znanstveni magistarski radovi Otvoreni pristup

Interpretacija dubokih klasifikatora i kvantitativne metrike za evaluaciju interpretacije

Poštić, Suraja

2026-03-16

Lokalna interpretacija klasifikatora slika dana je kao mapa istaknutosti čije vrijednosti odražavaju doprinose piksela slika predikciji klase. Ova teza predstavlja objektivnu proceduru za evaluaciju takvih algoritama. Dvije nove metode za interpretac…

Sveučilište u Zagrebu Sveučilište u Zagrebu Fakultet elektrotehnike i računarstva Zavod za elektroničke sustave i obradu informacija

Repozitorij FER-a

explainable artificial intelligence interpretation evaluation convolutional neural networks image classification model enhancement objašnjiva umjetna inteligencija evaluacija interpretacije konvolucijske neuronske mreže klasifikacija slika unaprjed¯enje modela

Prikaži detalje Otvori

Repozitorij FER-a

Zadnje dodano