hriaadn – Pagina 6 – HUB Român de Inteligență Artificială (HRIA)

feb.22026

Evaluating Latent Space Modeling in BiLSTM Autoencoders for Time Series Anomaly Detection

Evaluating Latent Space Modeling in BiLSTM Autoencoders for Time Series Anomaly Detection. Andrei-Alexandru Mariciuc, Lavinia Ferariu. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240246.

Aflați mai multe

feb.22026

Action Masking Methods for Safe Reinforcement Learning in a Non-Stationary Configurable Environment

Action Masking Methods for Safe Reinforcement Learning in a Non-Stationary Configurable Environment. Florin Leon. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240265.

Aflați mai multe

feb.22026

Generating and Optimizing What-If Scenarios Using a Transformers-Based Forecasting Model

Generating and Optimizing What-If Scenarios Using a Transformers-Based Forecasting Model. Alexandru Grigoraș, Florin Leon. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240464.

Aflați mai multe

feb.22026

Improving Maze Exploration and Evacuation with Reactive Agents through Directional Weighting

Improving Maze Exploration and Evacuation with Reactive Agents through Directional Weighting. Marius Gavrilescu, Gabriel-Alexandru Scînteie, Delia-Elena Bărbuță, Florin Leon. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240363.

Aflați mai multe

feb.22026

3D-CNNs Using Magnetoencephalography Spectral Data for Diagnosis of Parkinson’s Disease

3D-CNNs Using Magnetoencephalography Spectral Data for Diagnosis of Parkinson’s Disease. Robert-Cristian Bejan, Lavinia Ferariu. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240325.

Aflați mai multe

feb.22026

Semi-Supervised Learning Using Self-labelling and Consistency Regularisation for ECG Classification

Semi-Supervised Learning Using Self-labelling and Consistency Regularisation for ECG Classification. Mihai Aron, Lavinia Ferariu. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240516.

Aflați mai multe

feb.22026

Improving WAF Performance with Advanced ML Models: From RNN to GRU and LSTM

Improving WAF Performance with Advanced ML Models: From RNN to GRU and LSTM. Cristian Chindrus and Constantin-Florin Căruntu. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240342.

Aflați mai multe

feb.22026

Underbrake Fault Detection in Automotive Systems using Advanced Neural Network Architectures

Underbrake Fault Detection in Automotive Systems using Advanced Neural Network Architectures. Teofil-Constantin Ciobanu, Otilia-Cristiana Chioaru, George Maties, Letitia Mirea, Constantin-Florin Căruntu. In: Proceedings of the 29th International Conference on System Theory, Control and Computing, October 2025, Cluj-Napoca, România, DOI: 10.1109/icstcc66753.2025.11240553.

Aflați mai multe

feb.22026

Automotive NMOS Series Regulator with Frequency Compensation Based on Active Capacitance Multiplier that Improves Transient Response

Automotive NMOS Series Regulator with Frequency Compensation Based on Active Capacitance Multiplier that Improves Transient Response. Paul Coste, Paul Mireșan, Cristian Răducan, Marius Neag, Marina Țopa. In: Proceedings of the International Semiconductor Conference (CAS 2025), October 2025, Sinaia, Romania, DOI: 10.1109/cas66707.2025.11222332.

Aflați mai multe

feb.22026

A Machine Learning Working Methodology for Prediction and Failure Categorization of Power Supplies

A Machine Learning Working Methodology for Prediction and Failure Categorization of Power Supplies. Beatrix Pasztor, Alexandru Dumitru Buburuzan, Marius Purcar. In: Proceedings of the International Semiconductor Conference (CAS 2025), October 2025, Sinaia, Romania, DOI: 10.1109/cas66707.2025.11222344.

Aflați mai multe

feb.22026

Prose-to-P4: Leveraging High Level Languages

Prose-to-P4: Leveraging High Level Languages. Mihai Valentin Dumitru, Vlad Andrei Badoiu, Costin Raiciu. In: U.P.B. Scientific Bulletin, Series C, Vol. 87, Iss. 3, September 2025, ISSN: 2286-3540. (Link)

Aflați mai multe

feb.22026

Efficient Tissue Detection in Whole-Slide Images Using Classical and Hybrid Methods: Benchmark on TCGA Cancer Cohorts

Efficient Tissue Detection in Whole-Slide Images Using Classical and Hybrid Methods: Benchmark on TCGA Cancer Cohorts. Bogdan Ceachi, Filip Muresan, Mihai Trascau, Adina Magda Florea. In: Cancers Journal, Vol. 17, Iss. 17, September 2025, DOI: 10.3390/cancers17172918.

Aflați mai multe

feb.22026

Generating P4 Dataplanes Using LLMs

Generating P4 Dataplanes Using LLMs. Mihai-Valentin Dumitru, Vlad-Andrei Bădoiu, Alexandru M. Gherghescu, Costin Raiciu. In: Computer Networks Journal, Vol. 272, September 2025, DOI: 10.1109/hpsr62440.2024.10635926.

Aflați mai multe

feb.22026

Detection of Fake News in Romanian: LLM-Based Approaches to COVID-19 Misinformation

Detection of Fake News in Romanian: LLM-Based Approaches to COVID-19 Misinformation. Alexandru Dima, Ecaterina Ilis, Diana Florea and Mihai Dascalu. In: Information Journal, Vol. 16, Iss. 9, September 2025, DOI: 10.3390/info16090796.

Aflați mai multe

feb.22026

Hybrid Adaptive Greedy Algorithm Addressing the Multi-Robot Path Planning Problem

Hybrid Adaptive Greedy Algorithm Addressing the Multi-Robot Path Planning Problem. Aniko Kopacz Enol Garcia Gonzalez, Camelia Chira, Jose R. Villar. In: IEEE Latin America Transactions Journal, Vol. 23, Iss. 10, September 2025, DOI: 10.1109/TLA.2025.11150630.

Aflați mai multe

feb.22026

S-LAM3D: Segmentation-Guided Monocular 3D Object Detection via Feature Space Fusion

S-LAM3D: Segmentation-Guided Monocular 3D Object Detection via Feature Space Fusion. Diana-Alexandra Sas, Florin Oniga. In: Proceedings of the 2025 IEEE International Workshop on Multimedia Signal (MMSP), September 2025, Beijing, China, DOI: 10.1109/MMSP64401.2025.11324346.

Aflați mai multe

feb.22026

Panoptic segmentation techniques

Panoptic segmentation techniques. Elena-Claudia Maftei, Otilia Zvorișteanu. In: Proceedings of the 2025 IEEE International Workshop on Multimedia Signal Processing 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025) – Workshop on Remote Sensing, Computer Vision, and AI for a Sustainable Future (WSAI 2025), September 2025, Timișoara, România.

Aflați mai multe

feb.22026

Reaction-Diffusion Applications to Data Dynamic Clustering

Reaction-Diffusion Applications to Data Dynamic Clustering. Mihail-Cristian Heghea, Silviu-Dumitru Pavăl, Vasile-Ion Manta. In: Proceedings of the 2025 IEEE International Workshop on Multimedia Signal Processing 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025) – Workshop on Remote Sensing, Computer Vision, and AI for a Sustainable Future (WSAI 2025), September 2025, Timișoara,…

Aflați mai multe

feb.22026

Towards a Modular MultiGIS Architecture: Motivating Pilots in Drainage Design, Invasive Species Detection, and Heritage Analysis

Towards a Modular MultiGIS Architecture: Motivating Pilots in Drainage Design, Invasive Species Detection, and Heritage Analysis. Zhuoqian Wu, Andrei Ancuta, Maitane Urrutia Aparicio, Helga Hochbauer, Marc Frincu and Maryam Imani. In: Proceedings of 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025) – Workshop on Remote Sensing, Computer Vision, and AI…

Aflați mai multe

feb.22026

Analyzing and Forecasting the Evolution of Desertification in Southern Romania using Multispectral Imagery

Analyzing and Forecasting the Evolution of Desertification in Southern Romania using Multispectral Imagery. Rares-Bogdan Stefea, Andrei Scoferta and Marc Frincu. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025) – Workshop on Remote Sensing, Computer Vision, and AI for a Sustainable Future (WSAI 2025), September 2025, Timisoara,…

Aflați mai multe

feb.22026

Audio Deepfake Detection using MFCC Features and CNNs

Audio Deepfake Detection using MFCC Features and CNNs. Ioana-Victoria Zinveliu, Florica Dolca, Mircea Giurgiu. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

Objective Evaluation of Romanian TTS Systems: Glow-TTS and VITS

Objective Evaluation of Romanian TTS Systems: Glow-TTS and VITS. Oscar Gal, Mircea Giurgiu. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

Feature-Based AI/ML for ECG Artifact Detection: A Comparative Study of Supervised Classifiers

Feature-Based AI/ML for ECG Artifact Detection: A Comparative Study of Supervised Classifiers. Farago Paul, Laura Ioana Mihăilă, Claudia Georgiana Barbura, Sorin Hintea, Botond Sandor Kirei, Albert Fazakas. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

A Finetuned Llama-3 based Solution for Specific Information Retrieval with Enhanced Reliability

A Finetuned Llama-3 based Solution for Specific Information Retrieval with Enhanced Reliability. George Catalin Robotin, Radu Răzvan Slăvescu. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

Dissonant Ballerinas and Crafty Carrots: A Comparative Multi-modal Analysis of Italian Brain Rot

Dissonant Ballerinas and Crafty Carrots: A Comparative Multi-modal Analysis of Italian Brain Rot. Anca Dinu, Andra-Maria Florescu, Marius Micluța-Câmpeanu, Ștefana Arina Tăbușcă, Claudiu Creangă and Andreiana Mihail. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Cagliari, Italy.

Aflați mai multe

feb.22026

From signals to symbols: fault classification using transformers and logic-based reasoning

From signals to symbols: fault classification using transformers and logic-based reasoning. Darian Onchis, Eduard Hogea, Codruta Istin. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Cluj-Napoca, Romania.

Aflați mai multe

feb.22026

Explainability in AI-Driven Medical Diagnostics – (Invited Paper)

Explainability in AI-Driven Medical Diagnostics. Adina Magda Florea. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timișoara, România. (Invited Paper)

Aflați mai multe

feb.22026

Precision Matters: Comparing Deep Models for Visual Marker Detection on Eyewear-Mounted Calibration Objects

Precision Matters: Comparing Deep Models for Visual Marker Detection on Eyewear-Mounted Calibration Objects. Maduta Adrian. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

Salivary ferning segmentation with a lightweight CNN

Salivary ferning segmentation with a lightweight CNN. Ditmár Mate; Lehel Csató. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timisoara, Romania.

Aflați mai multe

feb.22026

Simple Idea Discovery in a Minimalist LLM Architecture Implementation

ContaGPT: A Domain-Adapted LLM for Romanian Financial and Accounting Applications. Cezar Tudor, Ștefan-Daniel Achirei, Cristian Simionescu. In: Proceedings of the 27th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC 2025), September 2025, Timișoara, România.

Aflați mai multe