About

researcher

Researcher at
Department of Computer Algorithms and Artificial Intelligence
University of Szeged.

Working on fully distributed machine learning algorithms.
Our gossip-based machie learning framework is available on github.

Thesis booklets (ENG/HUN)

Thesis slides (ENG/HUN)

Publications

Gábor Danner, István Hegedűs, Márk Jelasity. Improving Gossip Learning via Limited Model Merging. International Conference on Computational Collective Intelligence, ICCCI, 2023.
Balázs Nagy, István Hegedűs, Noémi Sándor, Balázs Egedi, Haaris Mehmood, Karthikeyan Saravanan, Gábor Lóki, Ákos Kiss. Privacy-preserving Federated Learning and its application to natural language processing. Knowledge-Based Systems, 2023.

Árpád Berta, Gábor Danner, István Hegedus, Márk Jelasity. Hiding Needles in a Haystack: Towards Constructing Neural Networks that Evade Verification. 2022 ACM Workshop on Information Hiding and Multimedia Security, IH&MMSEC'22, 2022.

Megyeri István, Hegedűs István, Jelasity Márk. Robust Classification Combined with Robust out-of-Distribution Detection: An Empirical Analysis. International Joint Conference on Neural Networks, IJCNN, 2021. (bib)

Danner Gábor, Hegedűs István, Jelasity Márk. Decentralized Machine Learning Using Compressed Push-Pull Averaging. Proceedings of the 1st International Workshop on Distributed Infrastructure for Common Good, 2020. (bib, doi)
Hegedűs István, Danner Gábor, Jelasity Márk. Decentralized learning works: An empirical comparison of gossip learning and federated learning. Journal of Parallel and Distributed Computing, 2020. (bib, doi)
Megyeri István, Hegedűs István, Jelasity Márk. Adversarial robustness of Model Sets. International Joint Conference on Neural Networks, IJCNN, 2020. (bib, doi)
Megyeri István, Hegedűs István, Jelasity Márk. Attacking model sets with adversarial examples. 28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN, 2020. (bib)

Hegedűs István, Danner Gábor, Jelasity Márk. Decentralized Recommendation based on Matrix Factorization: A Comparison of Gossip and Federated Learning. Decentralized Machine Learning at the Edge (DMLE'19 ECML-PKDD), 2019. (bib, doi)
Hegedűs István, Danner Gábor, Jelasity Márk. Gossip Learning as a Decentralized Alternative to Federated Learning. Distributed Applications and Interoperable Systems, 2019. (bib, doi)
Megyeri István, Hegedűs István, Jelasity Márk. Adversarial robustness of linear models: Regularization and dimensionality. 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN, 2019. (bib)

Danner Gábor, Berta Árpád, Hegedűs István, Jelasity Márk. Robust Fully Distributed Minibatch Gradient Descent with Privacy Preservation. Security and Communication Networks, 2018. (bib, doi)

Hegedűs István, Jelasity Márk. Distributed Differentially Private Stochastic Gradient Descent: An Empirical Study. 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP 2016), PDP'16, 2016. (bib, doi, slides)
Berta Árpád, Hegedűs István, Jelasity Márk. Dimension Reduction Methods for Collaborative Mobile Gossip Learning. 24th Euromicro International Conference on Parallel, Distributed, and Network-Based Processing (PDP 2016), PDP'16, 2016. (bib, doi, slides)
Hegedűs István, Berta Árpád, Jelasity Márk. Robust Decentralized Differentially Private Stochastic Gradient Descent. Journal of Wireless Mobile Networks, Ubiquitous Computing, and Dependable Applications (JoWUA), 2016. (bib)
Hegedűs István, Berta Árpád, Kocsis Levente, Benczúr A. András, Jelasity Márk. Robust Decentralized Low-Rank Matrix Decomposition. ACM Transactions on Intelligent Systems and Technology (TIST), 2016. (bib, doi)

Hegedűs István, Jelasity Márk, Kocsis Levente, Benczúr A. András. Fully distributed robust singular value decomposition. 14th IEEE International Conference on Peer-to-Peer Computing (P2P 2014), P2P'14, 2014. (bib, doi)
Berta Árpád, Hegedűs István, Ormándi Róbert. Lightning fast asynchronous distributed k-means clustering. 22th European Symposium on Artificial Neural Networks, ESANN 2014, 2014. (bib)

Hegedűs István, Ormándi Róbert, Jelasity Márk. Massively Distributed Concept Drift Handling in Large Networks. Advances in Complex Systems, ACS 16(4 & 5), 2013. (bib, doi)
Szörényi Balázs, Busa-Fekete Róbert, Hegedűs István, Ormándi Róbert, Jelasity Márk, Kégl Balázs. Gossip-based distributed stochastic bandit algorithms. International Conference on Machine Learning, ICML'13 28(3), 2013. (bib)

Hegedűs István, Nyers Lehel, Ormándi Róbert. Detecting Concept Drift in Fully Distributed Environments. IEEE 10th Jubilee International Symposium on Intelligent Systems and Informatics, SISY'12:183-188, 2012. (bib, doi, slides)
Hegedűs István, Ormándi Róbert, Jelasity Márk. Gossip-based Learning under Drifting Concepts in Fully Distributed Networks. 2012 IEEE Sixth International Conference on Self-Adaptive and Self-Organizing Systems, SASO'12:79 - 88, 2012. (bib, doi, slides)
Hegedűs István, Busa-Fekete Róbert, Ormándi Róbert, Jelasity Márk, Kégl Balázs. Peer-to-Peer Multi-Class Boosting. 18th International European Conference on Parallel and Distributed Computing (Euro-Par 2012), EuroPar'12:389 - 400, 2012. (bib, doi, slides)
Ormándi Róbert, Hegedűs István, Jelasity Márk. Gossip learning with linear models on fully distributed data. Concurrency and Computation: Practice and Experience, CPE, 2012. (bib, doi)

Ormándi Róbert, Hegedűs István, Jelasity Márk. Gossip Learning with Linear Models on Fully Distributed Data. (bib)
Ormándi Róbert, Hegedűs István, Jelasity Márk. Asynchronous Peer-to-peer Data Mining with Stochastic Gradient Descent. 17th International European Conference on Parallel and Distributed Computing (Euro-Par 2011), EuroPar'11:528 - 540, 2011. (bib, doi, slides)

Ormándi Róbert, Hegedűs István, Jelasity Márk. Overlay management for fully distributed user-based collaborative filtering. Proceedings of the 16th international Euro-Par conference on Parallel processing: Part I, EuroPar'10:446 - 457, 2010. (bib, doi, slides)
Ormándi Róbert, Hegedűs István, Farkas Richárd. Opinion mining by transformation-based domain adaptation. Proceedings of the 13th international conference on Text, speech and dialogue, TSD'10:157 - 164, 2010. (bib, doi, slides)
Ormándi Róbert, Hegedűs István, Csernai Kornél, Jelasity Márk. Towards Inferring Ratings from User Behavior in BitTorrent Communities. Proceedings of the 2010 19th IEEE International Workshops on Enabling Technologies: Infrastructures for Collaborative Enterprises, WETICE'10:217 - 222, 2010. (bib, doi, slides)
Farkas Richárd, Berend Gábor, Hegedűs István, Kárpáti András, Krich Balázs. Automatic free-text-tagging of online news archives. Proceeding of the 2010 conference on ECAI 2010: 19th European Conference on Artificial Intelligence, ECAI'10:529 - 534, 2010. (bib)
Hegedűs István, Ormándi Róbert, Farkas Richárd, Jelasity Márk. Novel Balanced Feature Representation for Wikipedia Vandalism Detection Task. Lab Report for PAN at CLEF 2010 - CLEF (Notebook Papers/LABs/Workshops), 2010. (bib, slides, results)

Farkas Richárd, Szarvas György, Hegedűs István, Almási Attila, Vincze Veronika, Ormándi Róbert, Busa-Fekete Róbert. Semi-automated Construction of Decision Rules to Predict Morbidities from Clinical Texts. Journal of the American Medical Informatics Association, 16(4):601 - 605, 2009. (bib, doi)