Improving scientific excellence and creativity in combating disinformation with artificial intelligence and language technologies

Informazioni relative al progetto

DisAI

ID dell’accordo di sovvenzione: 101079164

DOI

10.3030/101079164

Progetto chiuso

Data della firma CE 8 Luglio 2022

Data di avvio 1 Dicembre 2022

Data di completamento 30 Novembre 2025

Finanziato da

• Widening participation and spreading excellence

Costo totale

€ 1 499 750,00

Contributo UE

€ 1 499 750,00

Coordinato da

KEMPELENOV INSTITUT INTELIGENTNYCH TECHNOLOGII
Slovakia

CORDIS fornisce collegamenti ai risultati finali pubblici e alle pubblicazioni dei progetti ORIZZONTE.

I link ai risultati e alle pubblicazioni dei progetti del 7° PQ, così come i link ad alcuni tipi di risultati specifici come dataset e software, sono recuperati dinamicamente da .OpenAIRE .

Risultati finali

Data Management Plan (2)

Data management plan – Version 1

The first version of the data management plan (DMP) will be prepared by M6 to outline how the research data are collected or generated and will be handled during a project, and after it is completed, describe what data will be collected/generated and following what methodology and standards, whether and how this data will be shared and/or made open, and how it will be curated and preserved. The DMP is a living document: it will be updated by the end of each reporting period. DMP includes: (1) What data will be collected/generated; (2) What standards will be used; (3) What types and format of data will be created; (4) How will metadata be generated; (5) How will the data be documented; (6) What data will be exploited, shared, made open; (7) How and by whom will data be curated and preserved.

Data management plan – Version 2

The second updated version of the data management plan (DMP) will be prepared by M17 to outline how the research data is collected or generated by this point in the project and will be further handled during a project, and after it is completed, describe what data will be collected/generated and following what methodology and standards, whether and how this data will be shared and/or made open, and how it will be curated and preserved.

Documents, reports (4)

Plan for dissemination and exploitation including communication – Version 1

The document will outline the initial version of the detailed planning of the dissemination and communication activities in a systematic manner, with the aim of performing actions and campaigns that reach specific groups and audiences for specific purposes. The plan will also address the exploitation and sustainability aspects, ensuring the timely promotion of the project’s outcomes and engagement of parties outside the Consortium, interested to use or adopt them.

Report on dissemination, communication and synergy activities – Version 1

The report will outline the mid-term results of dissemination and communication activities in a systematic manner listing the performed actions and campaigns that reach specific groups and audiences for specific purposes. The report will also address the exploitation and sustainability mid-term results, focusing on the outcomes of promotion of the project’s outcomes and engagement of parties outside the Consortium, interested to use or adopt them.

Ethics, innovation and IPR management plan

The document will describe the planned measures for IPR management, as well as ethics and innovation management. The followed policies like the Open Science policy will be mentioned.

Replication challenge report

The report will describe the organization and results of the replication challenge targeted at doctoral students. Each involved student will be assigned a mentor from the leading partners. Together they will select a scientific paper (related to the project topic) and replicate the research described therein. The process and results of individual projects' realization will also be included in the report.

Data sets, microdata, etc (1)

Collected labelled dataset

The dataset with accompanying description including labelling scheme and selected quantitative characteristics. Data collection will include several approaches, e.g., automatic content crawling, crowdsourcing or using human experts to annotate data, based on requirements for various tasks. Data collected will be used to answer research questions posed in tasks 2.2-2.4 and will be treated following the FAIR principles.

Websites, patent fillings, videos etc. (1)

Visuals and branding materials

The plan will have a description of brand identity and visuals; website including a splash page; promotional material (brochure/leaflet, poster, roll-up, slides, poster, promo video); communication materials (press releases, newsletters, and additional mass mailing, “exploitation booster materials” to our target groups/stakeholders, aiming to contribute to its use and to maximize the impact of the project; social Media (SNS) management.

Pubblicazioni

Conference proceedings (41)

nvestigating Language and Retrieval Bias in Multilingual Previously Fact-Checked Claim Detection

Autori: Ivan Vykopal, Antonia Karamolegkou, Jaroslav Kopčan, Qiwei Peng, Tomáš Javůrek, Michal Gregor, Marián Šimko
Pubblicato in: 2025
Editore: Association for Computational Linguistics

o-MEGA: Optimized Methods for Explanation Generation and Analysis

Autori: Ľuboš Kriš, Jaroslav Kopčan, Qiwei Peng, Andrej Ridzik, Marcel Veselý, Martin Tamajka
Pubblicato in: 2025
Editore: Association for Computational Linguistics

HumanEval-XL: A Multilingual Code Generation Benchmark for Cross-lingual Natural Language Generalization

Autori: Peng, Qiwei; Chai, Yekun; Li, Xuhong
Pubblicato in: 2024
Editore: International Committee on Computational Linguistics
DOI: 10.48550/arXiv.2402.16694

Tokenization Falling Short: On Subword Robustness in Large Language Models

Autori: Chai, Yekun; Fang, Yewei; Peng, Qiwei; Li, Xuhong
Pubblicato in: Findings of the Association for Computational Linguistics: EMNLP 2024, 2024
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2406.11687

Large Language Models for Multilingual Previously Fact-Checked Claim Detection

Autori: Ivan Vykopal, Matúš Pikuliak, Simon Ostermann, Tatiana Anikina, Michal Gregor, Marian Simko
Pubblicato in: Findings of the Association for Computational Linguistics: EMNLP 2025, 2025
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2025.FINDINGS-EMNLP.852

Women Are Beautiful, Men Are Leaders: Gender Stereotypes in Machine Translation and Language Modeling

Autori: Matúš Pikuliak, Stefan Oresko, Andrea Hrckova, Marian Simko
Pubblicato in: Findings of the Association for Computational Linguistics: EMNLP 2024, 2024
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2024.FINDINGS-EMNLP.173

Concept Space Alignment in Multilingual LLMs

Autori: Peng, Qiwei; Søgaard, Anders
Pubblicato in: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 2024
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2410.01079

Multimodal and Multilingual Fact-Checked Article Retrieval

Autori: Stefanos-Iordanis Papadopoulos, Ivana Beňová, Sebastian Kula, Michal Gregor, George Karantaidis, Tomáš Javůrek, Marián Šimko, Symeon Papadopoulos
Pubblicato in: Proceedings of the 2025 International Conference on Multimedia Retrieval, 2025
Editore: ACM
DOI: 10.1145/3731715.3733402

GrEmLIn: A Repository of Green Baseline Embeddings for 87 Low-Resource Languages Injected with Multilingual Graph Knowledge

Autori: Daniil Gurgurov; Rishu Kumar; Simon Ostermann 0002
Pubblicato in: Findings of the Association for Computational Linguistics: NAACL 2025, 2025
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2409.18193

Comparing Specialised Small and General Large Language Models on Text Classification: 100 Labelled Samples to Achieve Break-Even Performance

Autori: Pecher, Branislav; Srba, Ivan; Bielikova, Maria
Pubblicato in: Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 2025
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2402.12819

Only for the Unseen Languages, Say the Llamas: On the Efficacy of Language Adapters for Cross-lingual Transfer in English-centric LLMs

Autori: Julian Schlenker, Jenny Kunz, Tatiana Anikina, Günter Neumann, Simon Ostermann
Pubblicato in: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop), 2025
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2025.ACL-SRW.62

Multilingual Previously Fact-Checked Claim Retrieval

Autori: Pikuliak, Matúš; Srba, Ivan; Moro, Robert; Hromadka, Timo; Smolen, Timotej; Melisek, Martin; Vykopal, Ivan; Simko, Jakub; Podrouzek, Juraj; Bielikova, Maria
Pubblicato in: 2023
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2305.07991

A Rigorous Evaluation of LLM Data Generation Strategies for Low-Resource Languages

Autori: Tatiana Anikina, Jan Cegin, Jakub Simko, Simon Ostermann
Pubblicato in: Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 2025
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2025.EMNLP-MAIN.418

SemEval-2025 Task 7: Multilingual and Crosslingual Fact-Checked Claim Retrieval

Autori: Qiwei Peng, Robert Moro, Michal Gregor, Ivan Srba, Simon Ostermann, Marian Simko, Juraj Podrouzek, Matúš Mesarčík, Jaroslav Kopčan, Anders Søgaard
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Understanding Subword Compositionality of Large Language Models

Autori: Qiwei Peng, Yekun Chai, Anders Søgaard
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Beyond Image-Text Matching: Verb Understanding in Multimodal Transformers Using Guided Masking

Autori: Ivana Benová; Jana Kosecka; Michal Gregor; Martin Tamajka; Marcel Veselý; Marián Simko
Pubblicato in: Lecture Notes in Computer Science ISBN: 9783031826696, 2025
DOI: 10.1007/978-3-031-82670-2_7

Credible, Unreliable or Leaked?: Evidence verification for enhanced automated fact checking

Autori: Chrysidis Z.; Papadopoulos S. I.; Papadopoulos S.; Petrantonakis P.
Editore: The Association for Computing Machinery
DOI: 10.1145/3643491.3660278

LLMs vs Established Text Augmentation Techniques for Classification: When do the Benefits Outweight the Costs?

Autori: Jan Cegin, Jakub Simko, Peter Brusilovsky
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Soft Language Prompts for Language Transfer

Autori: Ivan Vykopal; Simon Ostermann 0002; Marián Simko
Pubblicato in: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), 2025
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2407.02317

Assessing Web Search Credibility and Response Groundedness in Chat Assistants

Autori: Ivan Vykopal, Matúš Pikuliak, Simon Ostermann, Marián Šimko
Pubblicato in: Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics, 2025
Editore: Association for Computational Linguistics

skLEP: A Slovak General Language Understanding Benchmark

Autori: Marek Suppa, Andrej Ridzik, Daniel Hládek, Tomáš Javůrek, Viktória Ondrejová, Kristína Sásiková, Martin Tamajka, Marian Simko
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Pessimistic Off-Policy Optimization for Learning to Rank

Autori: Matej Cief, Branislav Kveton, Michal Kompan
Pubblicato in: Frontiers in Artificial Intelligence and Applications, ECAI 2024, 2024
Editore: IOS Press
DOI: 10.3233/FAIA240703

In-Depth Look at Word Filling Societal Bias Measures

Autori: Pikuliak, Matúš; Beňová, Ivana; Bachratý, Viktor
Pubblicato in: 2023
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2302.12640

Small Models, Big Impact: Efficient Corpus and Graph-Based Adaptation of Small Multilingual Language Models for Low-Resource Languages

Autori: Daniil Gurgurov, Ivan Vykopal, Josef Van Genabith, Simon Ostermann
Pubblicato in: 2025
Editore: Association for Computational Linguistics

On Multilingual Encoder Language Model Compression for Low-Resource Languages

Autori: Daniil Gurgurov, Michal Gregor, Josef Van Genabith, Simon Ostermann
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Debiasing Multilingual LLMs in Cross-lingual Latent Space

Autori: Qiwei Peng, Guimin Hu, Yekun Chai, Anders Søgaard
Pubblicato in: 2025
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2508.17948

Adapting Multilingual LLMs to Low-Resource Languages with Knowledge Graphs via Adapters

Autori: Daniil Gurgurov; Mareike Hartmann; Simon Ostermann 0002
Pubblicato in: Proceedings of the 1st Workshop on Knowledge Graphs and Large Language Models (KaLLM 2024), 2024
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2407.01406

Disinformation Capabilities of Large Language Models

Autori: Ivan Vykopal, Matúš Pikuliak, Ivan Srba, Robert Moro, Dominik Macko, Maria Bielikova
Pubblicato in: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2024
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2024.ACL-LONG.793

CV-Probes: Studying the interplay of lexical and world knowledge in visually grounded verb understanding

Autori: Benova, Ivana; Gregor, Michal; Gatt, Albert
Pubblicato in: CoRR, 2024
Editore: CogSci 2025
DOI: 10.48550/ARXIV.2409.01389

On Sensitivity of Learning with Limited Labelled Data to the Effects of Randomness: Impact of Interactions and Systematic Choices

Autori: Branislav Pecher; Ivan Srba; Mária Bieliková
Pubblicato in: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 2024
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2402.12817

Similarity Over Factuality: Are we Making Progress on Multimodal Out-of-Context Misinformation Detection?

Autori: Stefanos-Iordanis Papadopoulos, Christos Koutlis, Symeon Papadopoulos, Panagiotis C. Petrantonakis
Pubblicato in: 2025
Editore: IEEE/CVF

Cross-Validated Off-Policy Evaluation

Autori: Matej Cief; Branislav Kveton; Michal Kompan
Pubblicato in: Proceedings of the AAAI Conference on Artificial Intelligence, 2025
Editore: AAAI Press
DOI: 10.48550/ARXIV.2405.15332

Average Is Not Enough: Caveats of Multilingual Evaluation

Autori: Pikuliak, Matúš; Šimko, Marián
Pubblicato in: 2023
Editore: Association for Computational Linguistics
DOI: 10.48550/ARXIV.2301.01269

On Training Data Influence of GPT Models

Autori: Qingyi Liu; Yekun Chai; Shuohuan Wang; Yu Sun 0004; Qiwei Peng 0002; Keze Wang; Hua Wu 0003
Pubblicato in: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 2024
Editore: EMNLP 2024
DOI: 10.48550/ARXIV.2404.07840

dfkinit2b at CheckThat! 2025: Leveraging LLMs and Ensemble of Methods for Multilingual Claim Normalization

Autori: Tatiana Anikina, Ivan Vykopal, Sebastian Kula, Ravi Kiran Chikkala, Natalia Skachkova, Jing Yang, Veronika Solopova, Vera Schmitt, Simon Ostermann
Pubblicato in: 2025
Editore: CEUR

Use Random Selection for Now: Investigation of Few-Shot Selection Strategies in LLM-based Text Augmentation

Autori: Jan Cegin, Branislav Pecher, Jakub Simko, Ivan Srba, Maria Bielikova, Peter Brusilovsky
Pubblicato in: Findings of the Association for Computational Linguistics: EMNLP 2025, 2025
Editore: Association for Computational Linguistics
DOI: 10.18653/V1/2025.FINDINGS-EMNLP.296

Automatic Fact-checking in English and Telugu

Autori: Ravikiran Chikkala, Tatiana Anikina, Natalia Skachkova, Ivan Vykopal, Rodrigo Agerri, Josef van Genabith
Pubblicato in: 2025
Editore: INCOMA Ltd., Shoumen, Bulgaria

'Humor, Art, or Misinformation?': A Multimodal Dataset for Intent-Aware Synthetic Image Detection

Autori: Anastasios Skoularikis, Stefanos-Iordanis Papadopoulos, Symeon Papadopoulos, Panagiotis C. Petrantonakis
Pubblicato in: Proceedings of the 2nd International Workshop on Diffusion of Harmful Content on Online Web, 2025
Editore: ACM
DOI: 10.1145/3746275.3762215

Task Prompt Vectors: Effective Initialization Through Multi-task Soft Prompt Transfer

Autori: Róbert Belanec; Simon Ostermann 0002; Ivan Srba; Mária Bieliková
Pubblicato in: Lecture Notes in Computer Science ISBN 9783662722428, 2025
Editore: Springer-Verlag
DOI: 10.48550/ARXIV.2408.01119

Multilingual Political Views of Large Language Models: Identification and Steering

Autori: Daniil Gurgurov, Katharina Trinley, Ivan Vykopal, Josef Van Genabith, Simon Ostermann, Roberto Zamparelli
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Language Arithmetics: Towards Systematic Language Neuron Identification and Manipulation

Autori: Daniil Gurgurov, Katharina Trinley, Yusser Al Ghussin, Tanja Baeumel, Josef Van Genabith, Simon Ostermann
Pubblicato in: 2025
Editore: Association for Computational Linguistics

Peer reviewed articles (1)

ACM Computing Surveys

Autori: Branislav Pecher; Ivan Srba; Maria Bielikova
Pubblicato in: ACM Computing Surveys, 2024, ISSN 0360-0300
Editore: Association for Computing Machinary, Inc.
DOI: 10.48550/ARXIV.2312.01082