Gábor Recski
Univ.Ass. / PhD
About
Gábor Recski (he/him) is a computational linguist working in the field of natural language processing (NLP). Formerly an assistant professor at the Budapest University of Technology, he is currently a postdoctoral researcher in the Data Science Research Unit at the TU Wien Faculty of Informatics. His research interests include the computational modeling of natural language semantics and the use of such models for transparent information extraction. In addition to his research Gábor also teaches two courses on NLP at the Faculty and supervises several students at the BSc, MSc, and PhD levels. He is the co-author of over 50 peer-reviewed publications and a program committee member of ACL, EMNLP, LREC, as well as ACM CIKM and SIGIR.
Roles
- PostDoc Researcher
Data Science, E194-04 - Faculty Council
Substitute Member
Contact
- gabor.recski@tuwien.ac.at
- +43-1-58801-194412
- 1040 Wien, Favoritenstrasse 9 / Room HD0121
- vCard from TISS
- orcid.org/0000-0001-5551-3100
- informatics.tuwien.ac.at/people/gabor-recski
- tiss.tuwien.ac.at/person/336863
Courses
Projects
- 2026 / Datenkraftwerk GmbH
- 2019 – 2026 / European Commission / Publications (2)
- 2024 – 2025 / Gogolook Co Ltd
- 2024 – 2025 / Kontron Transportation GmbH
- 2022 – 2023 / Vienna Business Agency (WAW)
- 2022 / Siemens AG
- 2020 – 2021 / Botium GmbH
Publications
- Kovacs, A., Schmitt, P., Recski, G. (2025). KR Labs at ArchEHR-QA 2025: A Verbatim Approach for Evidence-Based Question Answering. In Proceedings of the 24th Workshop on Biomedical Language Processing (Shared Tasks) (pp. 69–74). Association for Computational Linguistics. Peer-reviewed.
- Arzt, V., Hanbury, A., Wiegand, M., Recski, G., Blevins, T. (2025). Relation Extraction or Pattern Matching? Unravelling the Generalisation Limits of Language Models for Biographical RE. In Proceedings of the 14th International Joint Conference on Natural Language Processing and the 4th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (pp. 1463–1484). Peer-reviewed.
- Recski, G., Iklodi, E., Lellmann, B., Kovács, Á., Hanbury, A. (2024). BRISE-plandok: a German legal corpus of building regulations. Language Resources and Evaluation. Peer-reviewed.DOI: 10.1007/s10579-024-09747-7 / Project: BRISE-Vienna
- Arzt, V., Azarbeik, M. M., Lasy, I., Kerl, T., Recski, G. (2024). TU Wien at SemEval-2024 Task 6: Unifying Model-Agnostic and Model-Aware Techniques for Hallucination Detection. In A. K. Ojha, A. S. Dogruöz, H. Tayyar Madabushi, G. Da San Martino, S. Rosenthal, A. Rosá (Eds.), Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024) (pp. 1183–1196). Association for Computational Linguistics. Peer-reviewed.
- Obereder, C., Recski, G. (2024). Word alignment in Discourse Representation Structure parsing. In Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024) (pp. 50–56). Peer-reviewed.HDL: 20.500.12708/201684
- Schmitt, P., Rakovics, Z., Rakovics, M., Recski, G. (2024). TPPMI - a Temporal Positive Pointwise Mutual Information Embedding of Words. In Proceedings of the 4th Workshop on Computational Linguistics for the Political and Social Sciences: Long and short papers (pp. 119–125). Peer-reviewed.HDL: 20.500.12708/201681
- Gemes, K. A., Kovacs, A., Recski, G. (2023). Offensive text detection across languages and datasets using rule-based and hybrid methods. In G. Drakopoulos E. Kafeza (Eds.), CIKM-WS 2022. Proceedings of the CIKM 2022 Workshops. CEUR-WS.org. Peer-reviewed.DOI: 10.34726/4341 / Download: PDF
- Recski, G., Kádár, F. (2023). Language complexity in human and machine translation: a preliminary study. In C. Orasan, R. Mitkov, G. Corpas Pastor, J. Monti (Eds.), International Conference on Human-Informed Translation and Interpreting Technology : HiT-IT 2023 : Proceedings (pp. 268–281). Incoma Ltd. Peer-reviewed.
- Kovacs, A., Gémes, K., Iklódi, E., Recski, G. (2022). POTATO: exPlainable infOrmation exTrAcTion framewOrk. In CIKM ’22: Proceedings of the 31st ACM International Conference on Information Knowledge Management (pp. 4897–4901). Association for Computing Machinery (ACM). Peer-reviewed.
- Kovacs, A., Gemes, K., Kornai, A., Recski, G. (2022). Explainable lexical entailment with semantic graphs. Natural Language Engineering, 1–24. Peer-reviewed.
- Csaky, R., Recski, G. (2021). The Gutenberg Dialogue Dataset. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. 16th Conference of the European Chapter of the Association for Computational Linguistics, Unknown. The Association for Computational Linguistics. Peer-reviewed.
- Recski, G., Lellmann, B., Kovacs, A., Hanbury, A. (2021). Explainable Rule Extraction via Semantic Graphs. In Proceedings of the Fifth Workshop on Automated Semantic Analysis of Information in Legal Text (ASAIL 2021) (pp. 24–35). CEUR-WS.org. Peer-reviewed.HDL: 20.500.12708/58465
- Gemes, K. A., Recski, G. (2021). TUW-Inf at GermEval2021: Rule-based and Hybrid Methods for Detecting Toxic, Engaging, and Fact-Claiming Comments. In Proceedings of the GermEval 2021 Workshop on the Identification of Toxic, Engaging, and Fact-Claiming Comments : 17th Conference on Natural Language Processing KONVENS 2021 (pp. 69–75). netlibrary.HDL: 20.500.12708/58499
- Bogensperger, J., Schlarb, S., Hanbury, A., Recski, G. (2021). DreamDrug - A crowdsourced NER dataset for detecting drugs in darknet markets. In Proceedings of the Seventh Workshop on Noisy User-generated Text (W-NUT 2021). 7th Workshop on Noisy User-generated Text (W-NUT 2021), Unknown. The Association for Computational Linguistics.
- Gemes, K. A., Kovacs, A., Reichel, M., Recski, G. (2021). Offensive text detection on English Twitter with deep learning models and rule-based systems. In P. Mehta, T. Mandl, P. Majumder, M. Mitra (Eds.), FIRE-WN 2021 [FIRE 2021 Working Notes] (pp. 283–296). CEUR-WS.org. Peer-reviewed.DOI: 10.34726/4342 / Download: PDF
- Recski, G., Kovacs, A., Gemes, K. A., Ács, J., Kornai, A. (2020). BME-TUW at SR’20: Lexical grammar induction for surface realization. In Proceedings of the Third Workshop on Multilingual Surface Realisation (MSR´20) (pp. 21–29). Peer-reviewed.HDL: 20.500.12708/55594
- Kovacs, A., Gemes, K., Kornai, A., Recski, G. (2020). Explainable lexical entailment with semantic graphs. In Proceedings of the 14th International Workshop on Semantic Evaluation (pp. 135–141). ACL Anthology. Peer-reviewed.HDL: 20.500.12708/58391
- Gemes, K. A., Kovacs, A., Recski, G. (2019). Machine comprehension using semantic graphs. In Proceedings of the Automation and Applied Computer Science Workshop 2019. AACS.HDL: 20.500.12708/58793
Presentations
- Recski, G. (2025, June 16). Transparent and trustworthy AI for legal document generation [Keynote Presentation]. 7th Workshop on Automated Semantic Analysis of Information in Legal Text (ASAIL 2025), United States of America (the). Invited.HDL: 20.500.12708/219130
- Recski, G. (2025, May 20). How can we trust LLMs? [Conference Presentation]. dataSTREAM 2025, Budapest, Hungary. Invited.HDL: 20.500.12708/219465
- Recski, G. (2024, November 19). Fact-checking LLMs with explainable information extraction [Conference Presentation]. Language Intelligence 2024, Austria. Invited.DOI: 10.34726/8540 / Download: PDF
- Recski, G. (2024, May 15). Nyelvi sokszínűség az emberi és gépi fordításban [Keynote Presentation]. Elektrubadúr plusz: Mesterséges intelligencia és műfordítás, Budapest, Hungary. Invited.HDL: 20.500.12708/199011 / Download: PDF
- Recski, G. (2024, February 16). What can AI do for Advanced Legal Research? [Conference Presentation]. IRIS24: Internationales Rechtsinformatik Symposion 2024, Salzburg, Austria. Invited.DOI: 10.34726/6140 / Download: PDF
- Recski, G. (2023, November 10). Natural Language Processing [Keynote Presentation]. Fachtagung Spracherkennung, Wien, Austria. Invited.HDL: 20.500.12708/190911
- Recski, G. (2023, November 8). What can AI do for Advanced Legal Research? [Keynote Presentation]. Law via the Internet Conference 2023, Wien, Austria. Invited.HDL: 20.500.12708/190952
- Gemes, K. A., Kovacs, A., Recski, G. (2022, October 21). Offensive Text Detection Across Languages and Datasets Using Rule-based and Hybrid Methods [Poster Presentation]. Advances in Interpretable Machine Learning and Artificial Intelligence (AIMLAI), Atlanta, US-GA, United States of America (the).DOI: 10.34726/3742 / Download: PDF
- Recski, G. (2022, June 22). Transparent information extraction from natural language [Presentation]. Complexity Science Hub talk, Austria. Invited.HDL: 20.500.12708/153675 / Download: PDF
- Recski, G. (2021). Explainable emotion detection with syntactic and semantic graphs. Österreichisches Treffen zu Sentimentinferenz (ÖTSI), Wien, Austria. Invited.HDL: 20.500.12708/87245
Theses
- Rathmayr, M. (2026). Latency-tiered hallucination detection : optimizing supervised-unsupervised pipelines for RAG systems [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2026.139744 / Download: PDF
- Kunerth, P. (2026). Evaluating extraction-based RAG : a systematic assessment of VerbatimRAG on the CLAPNQ benchmark [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2026.140323 / Download: PDF
- Beccard, L. (2026). Real-time prevention of factual hallucinations in retrieval-augmented generation [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2026.137481 / Download: PDF
- Selenge, T. (2026). Retrieval Augmented Generation: A Multi-Stage Architecture for Verbatim Financial Question Answering [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2026.140203 / Download: PDF
- Iszak, Z. (2025). Rule-based open information extraction from German legal domain [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.126169 / Download: PDF
- Bogenreiter, D. (2025). Explainable prediction of user post popularity : an analysis of the one million posts corpus [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.122385 / Download: PDF
- Sommer, M. (2025). Rule learning for open information extraction [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.122389 / Download: PDF
- Eder, C. (2025). Honeypot LLM : creation of the scam conversation corpus [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.122387 / Download: PDF
- Osmanaj, I. (2025). Open information extraction for fact-checking large language models [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.130140 / Download: PDF
- Ristic, K. (2025). Symbolic natural language inference for German open information extraction [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.130460 / Download: PDF
- Verdha, N. (2025). Multilingual hallucination detection for RAG applications [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.133020 / Download: PDF
- Csakvari, T. R. (2025). Large language model-based framework for open information extraction, triplet matching, and text comparison [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.131626 / Download: PDF
- Aleksić, L. (2025). Modelling inter-Individual differences in multimodal data : a metric learning approach for personalized well-being estimation in healthcare workers [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.133180 / Download: PDF
- Beham, T. (2024). Evaluating LIME-based explanations of relation extraction models [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2024.112944 / Download: PDF
- Hagmann, M. (2024). Extracting structured data from semi-structured computer screen specifications in German [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2024.117489 / Download: PDF
- Obereder, C. (2024). Aligning sentences to their formal meaning representation in the context of discourse representation structure parsing [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2024.120192 / Download: PDF
- Piwonka, P. (2024). Advanced pattern matching in graph-based relation extraction : a methodical approach to improving XAI NLP systems [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2024.120151 / Download: PDF
- Kopali, N. (2024). Cross-dataset medical entity recognition [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2025.121281 / Download: PDF
- Breiner, G. (2023). Graph working representations in hybrid models as explanations for common sense question answering [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2023.102461 / Download: PDF
- Kurteshi, M. (2023). Graph-based methods for user intent classification [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2023.105781 / Download: PDF
- Chytilek, F. (2023). Transforming text annotations into graph-based features for a human-in-the-loop explainable information extraction framework [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2023.112080 / Download: PDF
- Kaçuri, M. (2023). Explainability of hate speech classification for Albanian language using rule based systems and neural networks [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2023.105780 / Download: PDF
- Reichel, M. (2022). Explainability in hate speech detection [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2022.91421 / Download: PDF
- Bogensperger, J. (2021). Exploring transfer learning techniques for named Entity recognition in Nnoisy user-generated text [Diploma Thesis, Technische Universität Wien]. reposiTUm.DOI: 10.34726/hss.2021.86900 / Download: PDF